The
new engl a nd jour nal
o f
medicine
n engl j med 366;8 nejm.org february 23, 2012
697
original article
Colonoscopy versus Fecal Immunochemical
Testing in Colorectal-Cancer Screening
Enrique Quintero, M.D., Ph.D., Antoni Castells, M.D., Ph.D.,
Luis Bujanda, M.D., Ph.D., Joaquín Cubiella, M.D., Ph.D., Dolores Salas, M.D.,
Ángel Lanas,
M.D., Ph.D., Montserrat Andreu, M.D., Ph.D.,
Fernando Carballo, M.D., Ph.D., Juan Diego Morillas, M.D., Ph.D.,
Cristina Hernández, B.Sc., Rodrigo Jover, M.D., Ph.D., Isabel Montalvo, M.D., Ph.D.,
Juan Arenas, M.D., Ph.D., Eva Laredo, R.N., Vicent Hernández, M.D., Ph.D.,
Felipe Iglesias, R.N., Estela Cid, R.N., Raquel Zubizarreta, M.D., Teresa Sala, M.D.,
Marta Ponce, M.D., Mercedes Andrés, M.D., Gloria Teruel, M.D., Antonio Peris, M.D.,
María-Pilar Roncales, R.N., Mónica Polo-Tomás, M.D., Ph.D.,
Xavier Bessa, M.D., Ph.D., Olga Ferrer-Armengou, R.N., Jaume Grau, M.D.,
Anna Serradesanferm, R.N., Akiko Ono, M.D., José Cruzado, M.D.,
Francisco Pérez-Riquelme, M.D., Inmaculada Alonso-Abreu, M.D.,
Mariola de la Vega-Prieto, M.D., Juana Maria Reyes-Melian, M.D.,
Guillermo Cacho, M.D., José Díaz-Tasende, M.D., Alberto Herreros-de-Tejada, M.D.,
Carmen Poves, M.D., Cecilio Santander, M.D., and Andrés González-Navarro, M.D.,
for the COLONPREV Study Investigators*
The authors’ affiliations are listed in the
Appendix. Address reprint requests to Dr.
Castells at the Department of Gastroen-
terology, Hospital Clínic, Villarroel 170,
08036 Barcelona, Spain, or at castells@
clinic.ub.es; or to Dr. Quintero at the De-
partment of Gastroenterology, Hospital
Universitario de Canarias, Ctra. Ofra S/N
Cuesta, 38320 La Laguna, Tenerife, Spain,
or at equinter@gmail.com.
Drs. Quintero and Castells contributed
equally to this article.
*The investigators in the COLONPREV
study are listed in the Supplementary
Appendix, available at NEJM.org.
N Engl J Med 2012;366:697-706.
Copyright © 2012 Massachusetts Medical Society.
Abstr act
Background
Colonoscopy and fecal immunochemical testing (FIT) are accepted strategies for
colorectal-cancer screening in the average-risk population.
Methods
In this randomized, controlled trial involving asymptomatic adults 50 to 69 years of
age, we compared one-time colonoscopy in 26,703 subjects with FIT every 2 years in
26,599 subjects. The primary outcome was the rate of death from colorectal cancer
at 10 years. This interim report describes rates of participation, diagnostic findings, and
occurrence of major complications at completion of the baseline screening. Study
outcomes were analyzed in both intention-to-screen and as-screened populations.
Results
The rate of participation was higher in the FIT group than in the colonoscopy group
(34.2% vs. 24.6%, P<0.001). Colorectal cancer was found in 30 subjects (0.1%) in the
colonoscopy group and 33 subjects (0.1%) in the FIT group (odds ratio, 0.99; 95%
confidence interval [CI], 0.61 to 1.64; P = 0.99). Advanced adenomas were detected
in 514 subjects (1.9%) in the colonoscopy group and 231 subjects (0.9%) in the FIT
group (odds ratio, 2.30; 95% CI, 1.97 to 2.69; P<0.001), and nonadvanced adenomas
were detected in 1109 subjects (4.2%) in the colonoscopy group and 119 subjects
(0.4%) in the FIT group (odds ratio, 9.80; 95% CI, 8.10 to 11.85; P<0.001).
Conclusions
Subjects in the FIT group were more likely to participate in screening than were those
in the colonoscopy group. On the baseline screening examination, the numbers of
subjects in whom colorectal cancer was detected were similar in the two study
groups, but more adenomas were identified in the colonoscopy group. (Funded by
Instituto de Salud Carlos III and others; ClinicalTrials.gov number, NCT00906997.)
The New England Journal of Medicine
Downloaded from nejm.org at UNIVERSIDAD AUTONOMA DE MADRID RECTORADO on October 30, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour na l
o f
medicine
n engl j med 366;8 nejm.org february 23, 2012
698
C
olorectal cancer is the third most
common cancer worldwide and the second
leading cause of cancer-related deaths.
1
Several studies have shown that colorectal-cancer
screening is effective
2-5
and cost-effective
6
in the
average-risk population.
Recommended strategies for colorectal-cancer
screening fall into two broad categories: stool tests
(occult blood and exfoliated DNA tests) and struc-
tural examinations (flexible sigmoidoscopy, colo-
noscopy, and computed tomographic colonogra-
phy). Stool tests primarily detect cancer, and
structural examinations detect both cancer and
premalignant lesions.
2
Stool tests for occult blood
(guaiac testing and fecal immunochemical test-
ing [FIT]) are predominantly used in Europe and
Australia, whereas colonoscopy is the predomi-
nant screening method in the United States.
Colonoscopy is considered the most accurate
test for early detection and prevention of colorectal
cancer. Although data from randomized studies
evaluating the effect of colonoscopy on the rate of
death from colorectal cancer are lacking, the pro-
cedure is recommended as a first-line screening
test on the basis of indirect data and observational
studies. Population-based case–control studies
have suggested that colonoscopy markedly reduces
the risk of colorectal cancer
7,8
and death.
9
Recent
evidence suggests that patients with no abnor-
malities on a previous colonoscopy have a mark-
edly reduced risk of colorectal cancer.
8,10,11
In a
cohort of average-risk subjects, the use of screen-
ing colonoscopy was associated with a reduction in
the incidence of colorectal cancer of 67% and a
reduction in the rate of death of 65%.
12
Cohort
studies involving patients with adenomas have sug-
gested that polypectomy can prevent approximate-
ly 80% of colorectal cancers.
13,14
Comparative studies have shown that the semi-
quantitative FIT is more accurate than the guaiac
test for the detection of colorectal cancer and ad-
vanced adenomas,
15-19
and this new test is now
recommended as the first-choice fecal occult blood
test in colorectal-cancer screening. Although FIT
is less effective for neoplastic detection than colo-
noscopy or sigmoidoscopy, evidence suggests that
it may be better accepted,
20,21
and higher accep-
tance may counteract its lower detection capacity.
It has been suggested that FIT may be more effec-
tive and less costly than other screening strate-
gies. We conducted a randomized, controlled trial
to compare semiquantitative FIT with colonos-
copy. We hypothesized that FIT screening every
2 years would be noninferior to one-time colonosco-
py with respect to a reduction in mortality related
to colorectal cancer among average-risk subjects.
This interim report describes rates of participation,
diagnostic findings, and the occurrence of major
complications at the completion of the baseline
screening.
Methods
Study Design
We conducted this randomized, controlled, nonin-
feriority trial in eight Spanish regions (Aragón,
Basque Country, Canarias, Catalonia, Galicia, Ma-
drid, Murcia, and Valencia) with the participation of
15 tertiary care hospitals. The study was designed
to assess the efficacy of one-time colonoscopy and
biennial FIT for reducing the rate of death from
colorectal cancer at 10 years (primary trial out-
come). The study started in November 2008 with an
informative nationwide campaign.
22
The recruit-
ment period was initiated in June 2009, and the first
round finished in June 2011. Ten-year follow-up
will be completed in 2021.
The study protocol (available with the full text
of this article at NEJM.org) was approved by the
ethics committee at each hospital, and all subjects
provided written informed consent.
Study Population
Asymptomatic men and women between the ages
of 50 and 69 years were eligible for enrollment.
Exclusion criteria, which were ascertained after
randomization by means of a questionnaire at the
local screening office, included a personal history
of colorectal cancer, adenoma, or inflammatory
bowel disease; a family history of hereditary or
familial colorectal cancer (i.e., ≥2 first-degree rel-
atives with colorectal cancer or 1 in whom the dis-
ease was diagnosed before the age of 60 years)
23,24
;
a severe coexisting illness; and previous colecto-
my. Subjects were also temporarily excluded if
they had undergone fecal occult blood testing in
the previous 2 years or sigmoidoscopy or colo-
noscopy within the previous 5 years or if they
had symptoms requiring additional workup. The
subjects with previous screening tests became
eligible when sufficient time had elapsed since
the tests,
2
and those with symptoms became eli-
gible if the results of the clinical workup were
negative.
The New England Journal of Medicine
Downloaded from nejm.org at UNIVERSIDAD AUTONOMA DE MADRID RECTORADO on October 30, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
Colonoscopy vs. Fecal Immunochemical Testing
n engl j med 366;8 nejm.org february 23, 2012
699
Randomization
Subjects were identified through each Community
Health Registry, sorted according to household, and
stratified according to age (in 5-year age groups)
and sex. Households were randomly assigned in a
1:1 ratio to undergo either one-time colonoscopy
or biennial FIT. Randomization was performed be-
fore invitation with the use of a computer-generated
allocation algorithm on the basis of a randomized-
blocks method. Subjects were sent a preinvitation
letter containing information on colorectal-cancer
screening and the rationale for the study. Two
weeks later, an invitation letter was sent indicating
the subject’s study-group assignment. Two addi-
tional, reminder letters were mailed 3 and 6 months
after the invitation to subjects who did not respond
to the first mailed invitation. Subjects who agreed
to participate in the study received an appointment
at the local screening office, where they completed
the questionnaire. The study design allowed for
crossover between the two study groups.
Study Interventions
Among patients undergoing colonoscopy, bowel
cleansing and sedation were performed as de-
scribed previously.
25
All colonoscopies were per-
formed by experienced endoscopists (those who
had performed >200 colonoscopies per year).
26
Pol-
yps were categorized as non-neoplastic or neoplas-
tic. Adenomas measuring 10 mm or more in diam-
eter, with villous architecture, high-grade dysplasia,
or intramucosal carcinoma, were classified as ad-
vanced adenomas. Invasive cancer was considered
to be present when malignant cells were observed
beyond the muscularis mucosae. Advanced neo-
plasm was defined as advanced adenoma or inva-
sive cancer. Tumor staging, performed according
to the classification system of the American Joint
Committee on Cancer,
27
was based on the most
advanced lesion.
The FIT strategy consisted of analysis of a single
stool sample with the use of the automated semi-
quantitative OC-Sensor (Eiken Chemical) without
specific restrictions on diet or medication use.
Samples were processed as described previously
28
at each regional hospital. Subjects who were found
to have a hemoglobin level of 75 ng per milliliter
or more were invited to undergo colonoscopy.
Details regarding quality indicators for colonos-
copy are provided in the study protocol and in Ta-
ble 1 in the Supplementary Appendix, available at
NEJM.org.
Study Oversight
Palex Medical and Biogen Diagnóstica donated sup-
plies and automated fecal occult-blood analyzers
used for FIT but provided no other support for the
study. The companies were not involved in the de-
sign of the study, in the analysis or interpretation of
the data, or in the preparation of the manuscript.
Statistical Analysis
This study was based on the assumption that
screening average-risk subjects by means of bien-
nial FIT would not be inferior to one-time colonos-
copy with respect to the rate of death from colorec-
tal cancer at 10 years. The calculations were based
on an overall compliance rate of 30% and a crude
10-year rate of death from colorectal cancer of
6.96%.
29
Therefore, assuming a crude 10-year rate
of death from colorectal cancer of 1.74% among
subjects undergoing colonoscopy (a 75% reduction)
and of 3.41% among those screened by means of
FIT (a 51% reduction) and accepting a noninferi-
ority condition if the absolute difference was below
1.6 percentage points, we determined that a sample
of 55,498 subjects (27,749 in each study group)
would provide a power of 80%. A P value of less
than 0.025 was considered to indicate statistical
significance with the use of a one-sided test of
proportions.
30
We assessed study outcomes in both intention-
to-screen and as-screened analyses. In the latter
analysis, the detection rate was calculated as the
number of subjects with true positive results di-
vided by the number of subjects who actually un-
derwent testing. The diagnostic yield was the
number of subjects with true positive results
divided by the number of eligible subjects in the
intention-to-screen analysis. Subjects were exclud-
ed from the intention-to-screen analysis if they at-
tended the screening office visit and met one or
more exclusion criteria. Subjects who did not at-
tend the screening office visit and thus did not
provide information about exclusion criteria were
classified as eligible and were included in the
intention-to-screen analysis. Definitions of other
outcomes are provided in the study protocol. Be-
tween-group differences in rates of participation,
diagnostic yield, detection, and complications were
established by logistic-regression analysis, with
adjustment for age, sex, and participating center,
and are reported as odds ratios with 95% confi-
dence intervals. All analyses were performed with
the use of SPSS statistical software, version 15.0.
The New England Journal of Medicine
Downloaded from nejm.org at UNIVERSIDAD AUTONOMA DE MADRID RECTORADO on October 30, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour na l
o f
medicine
n engl j med 366;8 nejm.org february 23, 2012
700
Results
Study Population
Overall, 57,404 subjects were randomly assigned to
undergo either colonoscopy or FIT. Of these sub-
jects, 1970 could not be contacted and 2132 were
excluded either permanently (1.7% in the colonos-
copy group and 1.3% in the FIT group, P = 0.20) or
temporarily (2.2% in the colonoscopy group and
2.2% in the FIT group, P = 0.11) (Fig. 1). The eligible
population consisted of 26,703 subjects in the colo-
noscopy group and 26,599 in the FIT group. The
two groups were almost identical regarding both
mean (±SD) age (59.2±5.5 years in the colonoscopy
group and 59.3±5.6 years in the FIT group, P = 0.35)
and the proportion of subjects who were women
(53.5% in the colonoscopy group and 54.3% in the
FIT group, P = 0.25).
Participation
Among subjects who were assigned to undergo
colonoscopy, 5649 subjects accepted the pro-
posed strategy, whereas 1706 requested to be
screened by means of FIT (Fig. 1). Of the 5649
subjects who agreed to undergo colonoscopy,
4953 actually did so, and 1628 underwent FIT, for
a participation rate of 24.6%, according to the
intention-to-screen analysis (average age, 59.1±5.5
years; proportion of subjects who were women,
53.4%). Among subjects who were assigned to
undergo FIT, 9353 subjects accepted the pro-
posed strategy, whereas 117 asked to be screened
by colonoscopy. A total of 8983 subjects under-
went FIT, and 106 underwent colonoscopy, for an
overall participation rate of 34.2% (average age,
59.3±5.6 years; proportion of subjects who were
women, 54.4%). Therefore, there were differenc-
es between study groups regarding both the rate
of participation (odds ratio in the colonoscopy
group, 0.63; 95% confidence interval [CI], 0.60 to
0.65; P<0.001) and the crossover rate (odds ratio,
16.8; 95% CI, 13.9 to 20.2; P<0.001).
Diagnostic Yield
In the intention-to-screen analysis, colorectal can-
cer was detected in 30 subjects (0.1%) in the colo-
noscopy group and in 33 subjects (0.1%) in the FIT
group (odds ratio in the colonoscopy group, 0.99;
95% CI, 0.61 to 1.64; P = 0.99) (Table 1). Advanced
adenomas were found in 514 subjects (1.9%) in the
colonoscopy group and in 231 subjects (0.9%) in
the FIT group (odds ratio, 2.30; 95% CI, 1.97 to
2.69; P<0.001). Nonadvanced adenomas were found
in 1109 subjects (4.2%) in the colonoscopy group
and in 119 subjects (0.4%) in the FIT group (odds
ratio, 9.80; 95% CI, 8.10 to 11.85; P<0.001).
When the diagnostic yield was analyzed ac-
cording to the location of lesions, no significant
between-group difference was found for either
proximal or distal colorectal cancer (Table 2).
However, colonoscopy performed significantly bet-
ter than FIT in the diagnosis of advanced and
nonadvanced adenomas that were either proximal
or distal to the splenic flexure. The superior diag-
nostic yield of colonoscopy for advanced adenomas
was most evident for lesions in the proximal colon
(Table 2).
Detection Rate
On the basis of the screening that was actually per-
formed, 5059 subjects underwent colonoscopy and
10,611 underwent FIT (Fig. 1). Among subjects who
were screened by means of FIT, 767 (7.2%) tested
positive, and 663 of these subjects (86.4%) under-
went colonoscopy. Among subjects who were
screened by means of colonoscopy, 27 (0.5%) were
found to have colorectal cancer, as compared with
36 subjects (0.3%) who were screened by means of
FIT (odds ratio, 1.56; 95% CI, 0.93 to 2.56; P = 0.09)
(Table 3).
Tumor staging was similar in the two groups.
In the FIT group, 24 tumors were stage I, 6 were
stage II, and 6 were stage III. In the colonoscopy
group, 19 tumors were stage I, 6 were stage II, and
2 were stage III (P = 0.52). Colonoscopy was supe-
rior to FIT in the rates of detection of advanced
adenomas (odds ratio, 4.32; 95% CI, 3.69 to 5.07;
P<0.001) and nonadvanced adenomas (odds ratio,
25.98; 95% CI, 21.27 to 31.74; P<0.001) (Table 3).
No significant difference was observed in the
rate of detection of colorectal cancer when subjects
were stratified according to tumor location (Table
2 in the Supplementary Appendix). However, colo-
noscopy performed better than FIT with respect to
detection rates for advanced and nonadvanced ad-
enomas in both the proximal and distal colon.
Analysis of Resources
The numbers of subjects who needed to be screened
to find one colorectal cancer were 191 in the colo-
noscopy group and 281 in the FIT group, and the
numbers who needed to be screened to find any
advanced neoplasm were 10 and 36, respectively
(Table 3 in the Supplementary Appendix). However,
The New England Journal of Medicine
Downloaded from nejm.org at UNIVERSIDAD AUTONOMA DE MADRID RECTORADO on October 30, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
Colonoscopy vs. Fecal Immunochemical Testing
n engl j med 366;8 nejm.org february 23, 2012
701
57,404 Patients underwent randomization
28,708 Were assigned
to colonoscopy
28,696 Were assigned to FIT
916 Were not contacted 1054 Were not contacted
7368 Attended screening office 9512 Attended screening office
5649 Accepted colonoscopy
4953 Completed colonoscopy
92 Had positive results on FIT
76 Completed colonoscopy
675 Had positive results on FIT
587 Completed colonoscopy
Results:
4 CRC
32 Advanced adenomas
9 Nonadvanced adenomas
Results:
26 CRC
482 Advanced adenomas
1100 Nonadvanced adenomas
Results:
32 CRC
220 Advanced adenomas
103 Nonadvanced adenomas
Results:
1 CRC
11 Advanced adenomas
16 Nonadvanced adenomas
27,792 Were invited to undergo
colonoscopy
27,642 Were invited to undergo
FIT
472 Were permanently excluded
206 Had previous CRC or adenoma
56 Had inflammatory bowel disease
147 Had family history of CRC or polyposis
63 Had severe coexisting illness
617 Were temporarily excluded
60 Had colorectal symptoms
557 Had previous screening test
431 Were permanently excluded
170 Had previous CRC or adenoma
60 Had inflammatory bowel disease
177 Had family history of CRC or polyposis
24 Had severe coexisting illness
612 Were temporarily excluded
61 Had colorectal symptoms
551 Had previous screening test
26,703 Were eligible
for colonoscopy
26,599 Were eligible for FIT
117 Requested colonoscopy
106 Completed colonoscopy
1706 Requested FIT
1628 Completed FIT
9353 Accepted FIT
8983 Completed FIT
Figure 1. Enrollment and Outcomes.
A total of 1970 subjects were not contacted after being randomly assigned to undergo either colonoscopy or fecal immunochemical test-
ing (FIT) because they had died or had an inaccurate mailing address, which resulted in the return of the invitation letters. Criteria for
permanent exclusion were a personal history of inflammatory bowel disease, colorectal polyps, or colorectal cancer (CRC) and a family
history of CRC or polyposis syndromes. Temporary exclusion criteria were the presence of symptoms suggestive of colorectal disease
and occult blood testing within the previous 2 years or sigmoidoscopy or colonoscopy within the previous 5 years. The subjects with pre-
vious screening tests became eligible when sufficient time had elapsed since the tests, and those with symptoms became eligible if the
results of the clinical workup were negative.
The New England Journal of Medicine
Downloaded from nejm.org at UNIVERSIDAD AUTONOMA DE MADRID RECTORADO on October 30, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
