Differences in colorectal cancer survival between European and US
populations: the importance of sub-site and morphology§
G. Gattaa,*, L. Ciccolalloa, R. Capocacciab, M.P. Colemanc, T. Hakulinend, H. Møllere,
F. Berrinoa, and the EUROCARE Working Group
aIstituto Nazionale per lo Studio e la Cura dei Tumori, Division of Epidemiology, Via Venezian 1, 1-20133 Milan, Italy
bIstituto Supenore di Sanita ´, Department of Epidemiology and Biostatistics, Viale Regina Elena 299, 1-00161 Rome, Italy
cCancer and Public Health Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK
dFinnish Cancer Registry, Liisankatu 21B, FIN- 00170 Helsinki, Finland
eThames Cancer Registry, King’s College London, Capital House, 42 Weston Street, London SE1 3QD, UK
Received 17 February 2003; received in revised form 19 May 2003; accepted 16 June 2003
A previous study has shown a lower survival for colorectal cancer in Europe than in the United States of America (USA). It is of
interest to examine the extent to which anatomical location and morphological type influence this difference in colorectal cancer
survival. We analysed survival for 151244 European and 53884 US patients diagnosed with colorectal cancer aged 15–99 years
during the period of 1985–1989, obtained from 40 cancer registries that contribute to the EUROCARE study from 17 countries,
and nine Surveillance, Epidemiology and End-Results (SEER) registries in the USA. Cases included in the analysis were first pri-
mary malignant tumours (ICD-O behaviour code 3 or higher). Relative survival was estimated to correct for competing causes of
mortality. The Hakulinen–Tenkanen multiple regression approach was used to examine the prognostic impact of sub-site and ICD-
O histology codes. Relative excess risks (RERs) derived from this approach estimate the extent to which the hazard of death differs
from that in a reference region after adjustment for mortality in the general population. In order to explore geographical variation,
we defined three groups of European registries within which survival rates were known to be broadly similar. The proportion of
cases with unspecified sub-site was higher in Europe than the USA (10% versus 2%), but sub-site distributions were broadly similar
in the two populations. With the exception of appendix, 5-year survival was 13–22% higher in the USA than in Europe for each
anatomical sub-site. The proportion of non-microscopically-verified cases was higher in Europe than the USA (16 versus 3%).
Adenocarcinomas arising in a polyp (ICD-O-2 8210, 8261, 8263) were more frequent in the USA than Europe (13 versus 2%). Five-
year survival was higher in the USA than Europe for each morphological group, with the exception of non-microscopically-verified
cases. When age, gender and sub-site were considered, RERs ranged from 1.52 to 2.40 for the European populations (with the USA
as a reference). After inclusion of morphology codes, the range of RERs fell to between 1.28 and 1.86, mainly because of the high
frequency of adenocarcinoma in polyps in the USA. This analysis suggests that the large survival advantage for colorectal cancer
patients in the USA can only marginally be explained by differences in the distribution of sub-site and morphology. The main
explanatory difference is the proportion of adenocarcinoma in polyps.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: Relative survival; Population-based cancer registries; Colorectal cancer; Morphology; Sub-site; Europe; USA
Survival for most of the major adult cancers is higher
in the United States of America (USA) than in Europe,
especially among the oldest patients . In the USA, 5-
year relative survival for patients diagnosed with cancers
of the colon and rectum during 1985–1989 were 60 and
57%, respectively, while in Europe the figs. were 48% for
colon and 44% for rectum. These differences persisted in
0959-8049/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
European Journal of Cancer 39 (2003) 2214–2222
§Preliminary findings were presented at the 36th Annual Meeting
of the Groupe pour 1 ‘Enregistrement du cancer dans les pays de Lan
gue Latine (GRELL) in Neucha ˆ tel, Switzerland, 24–25 May 2001.
* Corresponding author. Tel.: +39-02-2390-3518; fax: +39-02-
E-mail addresses: email@example.com (G. Gatta), cicco-
firstname.lastname@example.org (L. Ciccolallo), email@example.com
(H. Møller), firstname.lastname@example.org (F. Berrino).
all 17 European populations studied, even the most afflu-
ent such as Sweden, Switzerland and The Netherlands.
In international comparisons of cancer survival, the
anatomical site of the malignancy is usually defined by
the three-digit code in the International Classification of
Diseases [2,3], and the morphologic type of the tumours
is rarely taken into account. However, survival from
some solid tumours is known to vary according to the
precise anatomical location (sub-site) within the organ
of origin, and by the morphological type of the tumour
[4,5]. The distribution of cancers by sub-site and mor-
phology also varies between countries .
The aim of this paper was to examine the extent to
which anatomical location and morphological type
influence the differences in colorectal cancer survival
between the USA and Europe.
2. Patients and methods
The European data were contributed by 40 popula-
tion-based cancer registries in 17 countries: four from
Northern Europe (Iceland, Finland, Sweden and Den-
mark), four from Eastern Europe (Slovenia, Slovakia,
Poland and Estonia), and nine from Western Europe
(Scotland, England, The Netherlands, Germany, Aus-
tria, Switzerland, France, Italy and Spain) as part of the
EUROCARE project [7,8]. The American data were
taken from the Surveillance, Epidemiology and End
Results (SEER) database, which is publicly available .
All patients diagnosed with cancer of the colon (ICD-
9 153) or rectum (ICD-9 154) aged 15–99 years during
the period of 1985–1989 were eligible for inclusion in
the analyses, with no selection for race. Only first, pri-
mary, malignant tumours (ICD-O-2 behaviour code 3
or higher) were eligible; in situ tumours and those of
uncertain or borderline malignancy were excluded. Both
histologically-verified and non-verified cases were inclu-
ded, but cases known to registries by death certificate
only (DCO) and those discovered incidentally at
autopsy were excluded.
More detailed information on the two databases is
available in two EUROCARE monographs [7,8] pub-
lished by the International Agency for Research on
Cancer, and in the periodic reports on the SEER pro-
gramme published under the auspices of the National
Cancer Institute [9–11].
The anatomical location of tumours was coded to the
Ninth Revision of the International Classification of
Diseases  (ICD-9), and morphology to the second
revision of the International Classification of Diseases
for Oncology  (ICD-O-2). Anatomical sub-sites were
defined by the fourth digit of ICD-9, and grouped into
four broader categories for analysis: right colon, left
colon, colon other and unspecified, and rectum (see
Table 2 footnote for detailed breakdown).
The distribution of colorectal cancer cases by mor-
phological group was broadly similar in European and
groupings based on those defined by Berg , were
used for the survival analysis: adenocarcinoma (11 sub-
groups), epidermoid carcinoma, other specified carci-
noma (4 sub-groups), sarcoma (2 sub-groups), other
specified morphology (2 sub-groups), unspecified mor-
phology and not microscopically-verified.
For seven European registries, and for the SEER data
as a whole, information on stage at diagnosis was
available in more than 75% of cases. Stage was classi-
fied into four categories: lesion confined to site of origin,
spread to immediately adjacent tissues and/or regional
lymph-nodes, spread to distant organs, and unknown
stage. The same stage classification was available for the
International comparisons of cancer survival require
adjustment for wide international differences in com-
peting causes of death (background mortality), to which
cancer patients are also subject [13,14]. Relative survival
is the ratio of the survival observed in the cancer
patients to the survival expected had they been subject
only to the mortality rates of the general population
from which they were drawn. Relative survival reflects
the excess mortality in cancer patients relative to back-
ground mortality. For univariate analysis, relative sur-
vival was estimated using the relevant life tables for
each population, specific for gender, calendar period
and year of age at death, using the Hakulinen method
. Regional life tables were needed for regional regis-
tries in Europe, and national life tables for the SEER
registries and the national registries in Europe. Multiple
regression analysis for grouped life table data  was
used to examine the prognostic impact of sub-site and
histology. The relative excess risk of death (RER) in Eur-
ope was estimated, with the USA as a reference category.
Since the European data refer to countries at widely
differing stages of economic development, and with dif-
ferent social structures and health care systems, we
defined three geographical groups of European regis-
tries within which survival rates were known to be
broadly similar . In Estonia, Poland, Slovakia and
Slovenia (‘Eastern Europe’), and in England, Scotland
and Denmark (United Kingdom (UK) and Denmark)
survival rates for colorectal cancer were lower than the
average for Europe. The other 10 participating Eur-
opean countries (Austria, France, Finland, Germany,
Iceland, Italy, The Netherlands, Spain, Sweden and
Switzerland) were combined.
The available data included 160381 cases of colo-
rectal cancer for Europe and 54471 cases for the USA.
G. Gatta et al./European Journal of Cancer 39 (2003) 2214–22222215
Autopsy-detected cases (0.4% in USA, 0.9% in Europe)
were excluded, as were DCO registrations (0.7% in USA,
4.8% in Europe). Total exclusions from the analysis were
9137 cases (5.7%) in Europe and 587 (1.1%) in USA. In
from Europe and 53884 from the USA (Table 1).
Less than 1% of cases were lost to follow-up the USA
(0.9%) or Europe (0.3%), although the proportion ran-
ged from 0 to 2% in the European registries.
In both Europe and the USA, approximately half the
cases were in men and half in women; the percentage of
men ranged from 45 to 56% among the European
countries. The overall proportion of patients aged 75
years or over at diagnosis was also similar in the USA
(36%) and Europe (38%), but geographical variation by
age was greater between the sub-sites. For colon cancer,
39% of cases in the USA were aged 75 years or over,
compared with 40% in Europe, but this proportion
varied widely between 45% (Switzerland) and 24% (Slo-
vakia). For rectal cancer, 30% of patients in the USA
were aged 75 years or over, compared to 35% in Europe;
again the latter proportion varied: between 40% (Swit-
zerland) and 25% (Estonia) (data not tabulated).
Cancers of the right colon were more common in the
USA than in Europe: 37 and 26% of the respective
totals. Among European countries, these figures ranged
from 37 to 18%. By contrast, rectal cancers were more
common in Europe: 30% of US cases and 39% of Eur-
opean cases (range 24–53%). Almost all USA cases
were microscopically-verified (97%) compared with an
average of 84% in Europe (range 63–99%), but in 10 of
the 11 Western European countries, the proportion was
similar to that of the USA (93–99%) (Table 1).
The sigmoid was the most common sub-site in the
colon in both series (42% Europe, 37% USA). The
proportion of cases with unspecified sub-site was con-
siderably higher in Europe (10%) than the USA (2%),
but sub-site distributions were broadly similar in the
two populations (Table 2).
Five-year survival was 13–22% higher in the USA
than in Europe for each of the anatomical sub-sites in
the colon and rectum (Table 2). The average 5-year
survival in Europe was less than 50% for each sub-site:
the only exception was for cancers of the appendix
(71% in Europe and 64% in the USA, based on 580
European and 290 US cases). Cancers at this sub-site
Numbers of cases contributed and characteristics (%) of study populations
on stage (%)
Europe (EUROCARE) 151244 50
53884 51 36
DCO, death certification only; SEER, Survelliance, Epidermiology and End-results.
Countries in bold are those for which national cancer registration data were available. Countries for which data from one or more regional cancer
registries are: Sweden (South Sweden); England (East Anglia, Merseyside and Cheshire, Oxford, South Thames, Wessex, West Midlands,Yorkshire);
Austria (Tyrol); France (Calvados, Co ˆ te d’Or, Doubs, Somme); Germany (Saarland); Switzerland (Basel, Geneva); The Netherlands (Eindhoven,
Rotterdam); Italy (Florence, Genoa, Latina, Modena, Parma, Ragusa, Romagna, Turin, Varese); Spain (Basque country, Mallorca, Navarra, Tar-
ragona); Poland (Cracow, Warsaw).
The numbers of cases cited are those included in the analyses. – Information on stage or morphology unavailable.
2216 G. Gatta et al./European Journal of Cancer 39 (2003) 2214–2222
have the highest survival, but comprise 1% or less of all
colorectal cancers in both populations. Overall survival
for cancer of the right colon was lower than cancer of
the left colon, both in the USA and Europe. Excluding
cancer of the appendix, 5-year survival for the ascending
colon was the highest of any sub-site in the right colon,
both in Europe (49%) and the USA (63%).
The proportion of cases for which no microscopic
verification was available was higher in Europe than the
USA (16% versus 3%) (Table 3). Of the micro-
scopically-verified cases, adenocarcinoma represented
96% of cases in the USA and 91% in Europe. The dif-
ference was greater for adenocarcinoma in polyp (13%
of cases in USA, 2% in Europe). With very few excep-
tions, 5-year survival was higher in the USA than in
Europe for each specific morphological group: for ade-
nocarcinoma in polyp (ICD-O-2 8210, 8261, 8263), sur-
vival rates were 89% in the USA and 80% in Europe.
Approximately 8% of colorectal cancers were mucinous
adenocarcinoma (ICD-O-2 8470, 8480, 8481, 8490)
slightly higher in the USA than Europe. As expected,
this morphological group had the lowest survival figures
Survival in Europe was higher than in the USA for
cases with unspecified topography (Colon, NOS), cases
with unspecified morphology (ICD-O-2 8000-8004) and
those that were not microscopically-verified.
The prognostic impact of sub-site and morphology on
the colorectal cancer survival differences between the
USA and Europe is shown in Table 4. In the simplest
model, adjusted for 1-year intervals of follow-up, the
relative excess risk (RER) of death was 2.34 in Eastern
Europe, 1.98 in the UK and Denmark, and 1.52 in the
other European countries. After further adjustment for
age and gender (model 2), the RER in Eastern Europe
increased slightly, because the population of cases there
was younger than average. Further adjustment for sub-
site (model 3) slightly reduced the RERs in Europe
because more European cases were colon NOS and rec-
tal cancers (at higher risk) than in the USA (see Table 2).
Further adjustment for morphology (model 4) reduced
the excess risks still further, to about 1.9 in Eastern
Europe, 1.4 in the UK and Denmark, and 1.3 in the
other European countries. This reflects the lower pro-
portion of adenocarcinomas in polyps (with a better
prognosis) and the higher proportion of tumours not
microscopically-verified in Europe compared with the
USA (see Table 3). All RERs were statistically sig-
nificant at the 5% level. We performed the same analy-
sis on the subset of cases for which sub-site, histology,
and microscopic verification were all available; we
observed results very close to those reported in Table 4:
RERs were 1.9 for Eastern Europe, 1.5 for UK and
Denmark, and 1.3 for other countries.
For the seven European registries for which informa-
tion on stage at diagnosis was available in more than
75% of cases, we also analysed for stage. Stage was
inserted into the model before sub-site and morphology.
Adjustment for stage only reduced the RERs (range
1.08–1.85; Finland and Slovenia, respectively) in all the
European registries considered, except Co ˆ te d’Or
(France). Minimal changes occurred when sub-site was
introduced. However, when morphology was included,
RERs reduced still further for Slovenia and England,
since these countries had low proportions of adeno-
carcinoma in polyps. By contrast, the RER for Switzer-
land increased from 1.22 to 1.29, as the Swiss data had
the highest proportion of adenocarcinoma in polyps.
Among these seven registries, Finland (29%) had the
highest and Co ˆ te d’Or (14%) the lowest percentages of
cases with distant metastasis; the figure was 19% for the
USA (data not tabulated). In addition, in this instance
we excluded the cases that were not microscopically
verified and those with missing information on subsite,
histology and stage, and reran the analysis. Again, the
RERs ranged between 1.6 for Co ˆ te d’Or, Varese and
Slovenia and 1.08 for Finland. The major difference was
found for Varese with 1.58 instead of 1.32, probably due
to the slight difference in the proportion of cases with
Colorectal cancer: number and proportion (%) of cases and 5-year
relative survival (%) by anatomical sub-site, Europe and USA,
patients diagnosed 1985–89
Cases Survival CasesSurvival
No. (%)% No. (%) %
580 (<1) 71
37506 (70) 62
19676 (37) 59
8233 (15) 59
16296 (30) 65
13581 (25) 67
16378 (30) 60
5570 (10) 62
9910 (18) 59
NOS, not otherwise specified.
Right colon: appendix (ICD-9 153.5), caecum (153.4), ascending colon
(153.6), hepatic flexure (153.0), transverse colon (153.1), splenic fiexure
(153.7). Left colon: descending colon (153.2), sigmoid colon (153.3).
‘Other colon’: tumours that overlap two or more fourth-digit sub-
categories within a three-digit rubric (153.8), colon unspecified (153.9).
Rectum: rectosigmoid junction (154.0), rectum (154.1), anus, including
anal canal (154.2), anus, unspecified (154.3) and other tumours of
rectum, rectosigmoid junction and anus (154.8).
G. Gatta et al./European Journal of Cancer 39 (2003) 2214–22222217
localised stage between the Varese Cancer Registry and
the USA (44.0 versus 41.0).
This study has analysed survival for colorectal cancer
in Europe and the USA taking into account age, sub-
site and morphology code. Stage was unavailable for the
majority of European registries so comparisons taking
stage into consideration was only performed between
the US and those European registries for which stage
information was available.
Nevertheless, a major staging indicator is incorpo-
rated into the morphology code adenocarcinoma in
polyps. Therefore, adjusting for morphology also par-
tially adjusts for stage.
Our results suggest that the large survival advantage
for colorectal cancer patients in the USA, compared
with those in Europe, can partly be explained by
differences in the distribution of morphological types,
but not by differences in the anatomical distribution of
tumours within the bowel. Our use of relative survival
took international differences in overall mortality into
account. Differences in the age distribution of cancer
patients had only a minor impact on the survival differ-
ences between Europe and the USA. Thus, inclusion of
age in the multiple regression analysis only had a mar-
ginal impact on the RERs, while stage at diagnosis
contributed to the observed differences. After adjusting
Coiorectal cancer: number and proportion (1%) of cases and five-year relative survival (1%) by morphology, Europe and USA, patients diagnosed
GroupaMorphology (ICD-O-2 M code range)EuropeUSA
Cases SurvivalCases Survival
No. (%)% No. (%)%
Adenocarcinoma in polyp/adenoma
Adenocarcinoma in adenomatous polyp (8210)
Adenocarcinoma in villous adenoma (8261)
Adenocarcinoma in tubulovillous adenoma (8263)
Cystoadenocarcinoma and mucinous adenocarcinoma (8470, 8480)
Mucin-producing adenocarcinoma (8481)
Signet ring cell carcinoma (8490)
Adenocarcinoma NOS (8140)
Papillary adenocarcinoma NOS (8260)
Villous adenocarcinoma (8262)
Adenocarcinoma in adenomatous polyposis coli (8220)
Other specified carcinomas
Small cell carcinoma (8041-5)
Undifferentiated carcinoma (8012-22, 8030-1, 8230-1, 8510)
Unspecified (carcinoma NOS)
Ill. Other specified types
IV. Unspecified (8000-4)
Total of microscopically-verified
V. No microscopic confirmationb
aNine morphological groups were used in the multiple regression analysis (see Table 4): 1 adenocarcinoma in polypladenoma; 2 mucinous ade-
nocarcinoma; 3 other adenocarcinoma, and adenocarcinoma NOS (reference group); 4 epidermoid carcinoma; 5 carcinoid; 6 other specified carci-
noma (except carcinoid), and carcinoma NOS; 7 sarcoma and other specified morphological types; 8 unspecified: 9 no microscopic confirmation.
NOS, not otherwise specified.
bPercentage of total number of cases. All other percentages relate to the total number of microscopically-verified cases.
2218 G. Gatta et al./European Journal of Cancer 39 (2003) 2214–2222
for age, gender, stage, sub-site, morphology (in that
order), the relative excess risk of death in Finland fell
from 45% to 8% (still significant), but the excess risks in
other parts of Europe ranged up to 61% (Slovenia).
This suggests that while the international differences in
survival are partly attributable to differences in the
nature of the disease in Europe and America, early
detection of colorectal cancer in the USA is also likely
to be a major determinant of the US survival advantage.
The patients included in this study were diagnosed
during the period 1985–1989, with followup to 1994.
The time-lag is partly due to the scale of the EURO-
CARE project, involving many cancer registries, some
with few resources, that had to ensure the completeness
and quality of their data for a given calendar period, with
at least 5 years of follow-up, before the analysis could
take place. Nevertheless, this study was designed to
explore the impact of differential case-mix on interna-
tional comparisons of colorectal cancer survival, and in
any case the US–Europe survival differential for patients
diagnosed during 1990–1994 is similar to that seen here
Because we used individual tumour records from the
SEER public-use dataset and the EUROCARE data-
base, it was possible to use the same analytical methods
for both datasets: the Hakulinen method for relative
survival , and the Hakulinen–Tenkanen approach
for multiple regression .
The main aspects of data quality that may affect the
analyses and conclusions are the proportion of cases
registered solely from a death certificate (DCO cases),
for which the survival time is unknown; the proportion
of cases for which microscopic verification (MV) was
available, and the proportion of cases that were lost to
follow-up. The main differences between the European
and US data were for DCO and MV cases. DCO cases
were more common in Europe than the USA (5% ver-
sus 1%). However, this difference cannot explain the
trans-Atlantic difference in survival, since DCO cases
generally have poorer survival than patients registered
during life  and if they could have been included, the
difference in survival would have been greater. The
considerably higher proportion of cases that were not
microscopically verified in Europe than the USA
(15.9% versus 2.8%) suggests that, for whatever reason,
fewer cancer patients receive surgery in Europe.
Another possible contributing factor may be that non-
surgical cases are less often registered by SEER regis-
tries than in Europe. Furthermore, the higher propor-
tion of European cases for which either sub-site or
histology were unspecified may be partly explained by a
reduced access to curative surgery.
Missing or unspecified data on detailed aspects of the
cancer diagnosis such as sub-site, morphology or stage,
characterised a fairly high proportion of European
cases. Generally, such unspecified cases have a much
poorer prognosis than average cases. This arises because
cases that present at a terminal stage often undergo less
intensive diagnostic work-up. We found that European
cases with unspecified or missing diagnostic data had
better survival than the corresponding, but much smal-
ler, group of US patients. Thus, we can consider the US
cases as a small subset of poor-prognosis patients, while
the much larger European subset may be a mixture of
cases with poor prognosis and those with average prog-
nosis. The systematic survival difference between cases
Relative excess risks of death, colorectal cancer, Europe versus USA, patients diagnosed 1985–1989: age, gender, sub-site, morphology and stage
Variable(s) added to previous model
Basic model Age and genderStage Sub-site Morphology
Age, gender, sub-site and morphology
UK and Denmark
Other European countries
Age, gender, sub-site, morphology and stage
Co ˆ te d’Or
Basel and Geneva
USA is reference category. Eastern Europe: Estonia, Poland, Slovakia, Slovenia; UK and Denmark: England, Scotland, Denmark; Other European
countries: Austria, Finland, France, Germany, Iceland, Italy, The Netherlands, Spain, Sweden and Switzerland.
Levels of categorical variables included in models; three age groups: 15–54 years, 55–74 years, 75 years and over; gender: males and females; four
sub-sites: right colon, left colon, other colon, rectum; nine morphological groups (see Table 3); four stage categories: tumour confined to the site of
origin, tumour spread to immediately adjacent tissues and/or regional lymph-nodes, tumour with metastases in distant organs, no information on
stage or stage not available.
G. Gatta et al./European Journal of Cancer 39 (2003) 2214–22222219
with unspecified diagnostic information is important
when these cases are included as a separate category in
the multivariate analysis. However, exclusion of all such
cases did not materially influence the analysis presented
in Table 4.
Five European registries (Co ˆ te d’Or, Finland, Oxford,
Saarland and Sweden) did not classify any tumours as
adenocarcinoma in polyps (ICD-O-2 8210, 8261, 8263),
coding them instead simply as adenocarcinoma (8140).
Survival for adenocarcinomas arising in a polyp in
Europe was substantially higher (80%) than for unspe-
cified adenocarcinoma (47%), and erroneous inclusion
of these tumours among the unspecified adenocarcino-
mas could have inflated survival for the latter category
in Europe. The proportion of adenocarcinoma in polyps
among European cases increased to 5.3% after exclu-
sion of these registries (cf. 13% in the USA), but the
5-year survival rate for adenocarcinoma NOS barely
changed as a result (46% cf. 47%). Most adenocarci-
nomas in polyps (77%) occurred in the left colon and
rectum, and 80% were of localised stage. The two mor-
phological categories were analysed separately in the
multivariate analysis, but inclusion of sub-site and stage
should have reduced any effect of misclassification.
Race was not considered in this study. Data on race
are not collected at cancer registration in Europe. The
proportion of non-whites is lower in the EUROCARE
than SEER populations: approximately 10% of the US
cancer cases were black. White patients have a higher
than average survival in the USA , so if the analysis
had been restricted to US whites, the US–Europe survi-
val differences would have been even larger. For colo-
rectal cancer, race should probably be considered more
as a proxy of social class than an indication of a differ-
ent type of disease.
Colorectal cancers are characterised by a much better
response when treated at an early stage, and the large
survival differences may therefore reflect the fact that
more healthy Americans than Europeans undergo early
diagnostic procedures. One indicator of early diagnosis
is the proportion of all colorectal cancers that are
adenocarcinomas in polyps, and this figure was much
higher among US cases than in Europe (13% versus
2%). A survey by the National Health Interview Survey
(NHIS) and the Behavioural Risk Factor Surveillance
System (BRFSS)  suggests that the prevalence of
colorectal cancer screening in the USA is high. In 1987,
the proportion of healthy people aged 50 years or over
who reported undergoing proctoscopy or sigmoido-
scopy at some time in the past varied between 24 and
37% by age. In the same population, the proportion
who reported having had a faecal occult blood test
(FOBT) in the preceding year ranged from 16 to 24%.
We do not have equivalent information for Europe, but
access to screening or early diagnostic procedures is
certainly much lower. The proportion of adenocarci-
noma in polyps ranged from 16% for the Swiss regis-
tries to less than 1% in Austria, Poland, Slovenia and
the UK, which suggests that early diagnostic procedures
are much less widely available in Europe than in the
parts of the USA covered by the SEER programme.
Geographical variation in colorectal cancer survival
was much greater in Europe than the USA. Five-year
survival was very low in Eastern Europe (below 25% in
Poland, and no more than 35% in Estonia, Poland and
Slovenia; while in Western Europe, survival varied from
41% to 54% for colon cancer and from 38 to 53% for
rectal cancer . Among patients diagnosed in the nine
SEER registry areas during the same period (1985–
1989), colorectal cancer survival ranged from 56 to 65%
. Differences in the availability of early diagnostic
procedures and in access to effective treatment are likely
to be the main causes of European survival differences.
High survival rates in the USA, with little geographical
variation between the SEER registries, could arise from
the fact that the SEER registries are not fully repre-
sentative of the entire US population, as they cover the
more affluent areas of the USA . Broadening the
comparison to include a wider range of US populations
would be an interesting future step.
Stage is a crucial prognostic factor to be considered in
explaining colorectal cancer
Adjustment for stage in this study, even in broad cate-
gories, markedly reduced the differences in survival
between some European registries and the USA.
Unfortunately, stage information was not available for
all cases collected by European cancer registries, and the
available data were not always directly comparable. A
major goal of future studies is to obtain comparable
information on stage from large representative samples
of cancer patients in each of the contributing cancer
registries: this will pinpoint remediable causes of inter-
national cancer survival differences and make it possible
to quantify the impact of removing them.
5. The EUROCARE Working Group for this study
Austria: W. Oberaigner (Cancer Registry of Tyrol);
Denmark: H.H. Storm (Danish Cancer Society); Esto-
nia: T. Aareleid (Estonian Cancer Registry); Finland: T.
Hakulinen (Finnish Cancer Registry); France: H.
Lefevre (Calvados Digestive Cancer Registry), J. Mace-
Lesec’h (Calvados General Cancer Registry), P. Arveux
(Doubs Cancer Registry), H. Mathieu-Daude’ (Herault
Cancer Registry) N. Raverdy (Somme Cancer Registry);
Germany: H. Ziegler (Saarland Cancer Registry); Ice-
land: L. Tryggvadottir (Icelandic Cancer Registry);
Italy: P. Crosignani, (Lombardy Cancer Registry), S.
Ferretti (Ferrara Cancer Registry), E. Conti (Latina
Cancer Registry), M. Vercelli, A. Quaglia (Liguria
Cancer Registry, NCI, Genova), M. Federico, L. Man-
2220G. Gatta et al./European Journal of Cancer 39 (2003) 2214–2222
gone (Modena Cancer Registry), M. Ponz De Leon
(Modena Colorectal Cancer Registry), V. De Lisi, L.
Serventi (Parma Cancer Registry), R. Zanetti (Piedmont
Cancer Registry), L. Gafa ` , R. Tumino (Ragusa Cancer
Registry), F. Falcini (Romagna Cancer Registry), E.
Crocetti, E. Paci, (Tuscan Cancer Registry), Poland: J.
Rachtan, (Cracow Cancer Registry), M. Bielska-Lasota,
Z. Wronkowski (Warsaw Cancer Registry); Slovakia: A.
Obsitnikova, I. Plesko, (National Cancer Registry of
Slovakia); Slovenia: V. Pompe-Kirn (Cancer Registry of
Slovenia); Spain: C. Martinez-Garcia (Granada Cancer
Registry), I. Garau (Mallorca Cancer Registry), C.
Navarro (Murcia Cancer Registry) E. Ardanaz, C.
Moreno (Navarra Cancer Registry), J. Borra ´ s, J. Gal-
ceran (Tarragona Cancer Registry); Sweden: T. Mo ¨ ller
(Southern Swedish Regional Tumour Registry); Swit-
zerland: J. Torhorst (Basel Cancer Registry), J.M. Lutz,
C. Bouchardy (Geneva Cancer Registry); The Nether-
lands: J.W.W. Coebergh (Eindhoven Cancer Registry),
R.A.M. Damhuis (Rotterdam Cancer Registry); Scot-
land: R. Black, (Scottish Cancer Intelligence Unit);
United Kingdom: T.W. Davies, S. Godward (East
Anglian Cancer Registry), E.M.I. Williams (The Mer-
seyside and Cheshire Cancer Registry), D. Forman
(Northern and Yorkshire Cancer Registry and Infor-
mation Service & Centre for Cancer Research), M.
Roche, S. Edwards (Oxford Cancer Intelligence Unit),
J. Smith (South and West Cancer Intelligence Unit), H.
Møller, J. Bell (Thames Cancer Registry), G. Lawrence
(West Midlands Cancer Intelligence Unit).
5.1. Steering Committee and Data Analysis Centre
F. Berrino (Project Leader), A. Micheli, M. Sant
(Epidemiology Unit, Istituto Nazionale per lo Studio e
la Cura dei Tumori, Milan, Italy); A. Verdecchia,
(National Institute of Health, Rome, Italy); J.W.W.
Coebergh (The Netherlands); M.I.P. Coleman (UK); J.
Este ´ ve (Centre Hospitalier Lyon Sud, France); J. Faivre
(France); T. Hakulinen (Finland); C. Martinez-Garcia
(Spain); H. Møller (UK).
This work would not have been possible without the
sustained effort over many years of cancer registries
across Europe, and we are extremely grateful for their
co-operation. The EUROCARE study was financed
through the BIOMED Programme of the European
Community. This study was also financed by the ‘Eur-
ope against Cancer’ programme (project number S
12.117414 ‘Differences in the survival of colorectal can-
cer patients between Europe and USA’). The authors
are grateful to Emily Taussig for editorial assistance and
Donald Ward for help with the English.
1. Gatta G, Capocaccia R, Coleman MP, et al. Toward a compar-
ison of survival in American and European cancer patients.
Cancer 2000, 89, 893–900.
2. World Health Organisation. International Classification of Dis-
eases for Oncology (ICD-O), 2nd edn. Geneva, WHO, 1990.
3. World Health Organisation. International Classification of Dis-
eases, 1975, 9th edn. Geneva, WHO, 1997.
4. Thomas RM, Sobin LH. Gastrointestinal cancer. Cancer 1995,
5. Berrino F, Gatta G. Variation in survival of patients with head
and neck cancer in Europe by the site of origin of the tumours.
Eur J Cancer 1998, 34, 2154–2161.
6. Parkin DW, Whelan SL, Ferlay J, Raymond L, Young J. Cancer
Incidence in Five Continents, Vol. VII. ARC Scientific Publications
No. 143. Lyon, International Agency for Research on Cancer,
7. Berrino F, Capocaccia R, Gatta G, et al, eds. Survival of Cancer
Patients in Europe: The EUROCARE 2-Study. IARC Scientific
Publications No. 151. Lyon, International Agency for Research
on Cancer, 1999.
8. Berrino F, Sant M, Verdecchia A, Capocaccia R, Hakulinen T,
Este ´ ve J, eds. Survival of Cancer Patients in Europe: The EURO-
CARE Study. IARC Scientific Publications No. 132. Lyon, Inter-
national Agency for Research on Cancer, 1995.
9. US Department of Health and Human Services. National Cancer
Institute, Surveillance, Epidemiology and End Results Program.
CDROM]. Bethesda, MD, US Department of Health and
Human Services, 2000.
10. Miller BA, Ries LAG, Hankey BF, Kosary CL, Edwards BK.
Cancer Statistics Review: 1973–1989. NTH Publication No. 92-
2789. Bethesda, MD, National Cancer Institute, 1992.
11. Kosary CL, Ries LA, Miller BA, Hankey BF, Harras A,
Edwards BK. SEER Cancer Statistics Review, 1973–1992: Tables
and Graphs. NIH Publication No. 96-2789. Bethesda, MD,
National Cancer Institute, 1995.
12. Berg JM. Morphological classification of human cancer. In
Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and
Prevention. New York, OUP, 1996, 28–44.
13. Micheli A, De Angelis R, Giorgi Rossi A, Capocaccia R. General
mortality and the survival of cancer patients in Europe. In Ber-
rino F, Capocaccia R, Gatta G, et al., eds. Survival of Cancer
Patients in Europe: the EUROCARE-2 Study. IARC Scientific
Publications No. 151. Lyon, International Agency for Research
on Cancer, 1999.
14. Hakulinen T. Cancer survival corrected for heterogeneity in
patient withdrawal. Biometrics 1982, 38, 933–942.
15. Hakulinen T, Abeywickrama KH. A computer program package
for relative survival analysis. Comput Prog Biomed 1985, 19, 197–
16. Hakulinen T, Tenkanen L. Regression analysis of relative survi-
val rates. Appl Stat 1987, 19, 197–207.
17. Sant M, Capocaccia R, Coleman MP, et al. Cancer survival
increases in Europe, but international differences remain wide.
Eur J Cancer 2001, 37, 1659–1667.
18. Berrino F, Capocaccia R, Gatta G, et al. Survival of cancer
patients in Europe: the EUROCARE 3-study. Ann Oncol, [in
19. Gatta G, Capocaccia R, Sant M, et al. Understanding variations
in survival for colorectal cancer in Europe: a EUROCARE high
resolution study. Gut 2000, 47, 533–538.
20. Berrino F, Este ´ ve J, Coleman MP. Basic issues in estimating and
comparing the survival of cancer patients. In Berrino F, Sant M,
Verdecchia A, Capocaccia R, Hakulinen T, Este ´ ve J, eds. Survival
G. Gatta et al./European Journal of Cancer 39 (2003) 2214–22222221
of cancer patients in Europe: The EUROCARE Study. IARC Sci-
entific Publications No. 132. Lyon, International Agency for
Research on Cancer, 1995, 1–14.
21. Mayberry RM, Coates RI, Hill HA, et al. Determinants of black/
white differences in colon cancer survival. J Natl Cancer Inst
1995, 87, 1686–1693.
22. Ries LA, Wingo PA, Miller DS, et al. The annual report to the
nation on the status of cancer, 1973–1997, with a special section
on colorectal cancer. Cancer 2000, 88, 2398–2424.
23. Mariotto A, Capocaccia R, Verdecchia A, et al. Projecting SEER
cancer survival rates to the US: an ecological regression
approach. Cancer Causes Control 2002, 13, 101–111.
2222G. Gatta et al./European Journal of Cancer 39 (2003) 2214–2222