Journal of Clinical Epidemiology 66 (2013) 202e208

Matching of controls may lead to biased estimates of specificity

in the evaluation of cancer screening tests
Hermann Brennera,*, Lutz Altenhofenb, Sha Taoa
a
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, INF 581, D-69120 Heidelberg, Germany
b
Central Research Institute of Ambulatory Health Care in Germany, Hervert-Lewin-Platz 3, D-10623 Berlin, Germany
Accepted 24 September 2012

Abstract
Objectives: In the evaluation of cancer screening tests, cancer-free controls are often matched to cancer cases on factors such as sex and
age. We assessed the potential merits and pitfalls of such matching using an example from colorectal cancer (CRC) screening.
Study Design and Setting: We compared sex and age distribution of CRC cases and cancer-free people undergoing screening colono-
scopy in Germany in 2006 and 2007. We assessed specificity by sex and age of two immunochemical fecal occult blood tests (iFOBTs) in
a study among screening colonoscopy participants conducted in the same years, and we assessed the expected impact of matching by sex
and age on the validity of specificity estimates at various cut points.
Results: In the screening colonoscopy program, the proportion of men and mean age were 59.6% and 68.6 years among 10,324 CRC
patients compared with 45.6% and 64.7 years, respectively, among 997,490 cancer-free participants. The specificity of the iFOBTs was
higher among women than among men and decreased with age. Matching of cancer-free controls by age and sex would have led to the
underestimation of specificity at all cut points assessed.
Conclusion: In the evaluation of cancer screening tests, matching of controls may lead to biased estimates of specificity. Ó 2013
Elsevier Inc. All rights reserved.
Keywords: Bias; Early detection; Matching; Screening; Specificity; Statistical methods

1. Introduction in screening populations, cancer-free convenience samples

In the ‘‘omics’’ era, research on novel early detection
markers for cancer is blooming [1,2]. In the evaluation of
cancer early detection markers, cancer patients are re-
cruited for determining sensitivity. For the evaluation of
specificity, noncancer controls are needed. Ideally, cancer-
free controls should be recruited from the screening popu-
lation, but, given the difficulty of verification of the absence
of cancer by a (often rather invasive) gold standard method
Conflict of interest statement: None.
Competing interests: There are no competing interests.
Grant support: The German screening colonoscopy registry is funded
by the National Association of Statutory Health Insurance Physicians
and the National Health Insurance Fund. The BLITZ study was supported
in part by the German Research Foundation (Deutsche Forschungsgemein-
schaft) within the framework of a PhD program (Graduiertenkolleg 793)
and by a grant from the German Federal Ministry of Education and Re-
search (O1ESO72). The test kits were provided free of charge by the man-
ufacturer. The sponsor did not have any role in this study.
* Corresponding author. Tel.: þ49-6221-421301; fax: þ49-6221-
421302.
E-mail address: h.brenner@dkfz.de (H. Brenner).
0895-4356/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jclinepi.2012.09.008
recruited in clinical settings are often used in practice.
The suitability of controls is often judged by their compa-
rability with cases with respect to the distribution of key
sociodemographic factors, such as sex and age. In some
studies, matching by these factors is used to ensure full
comparability. For example, in a recent systematic review
on the performance of blood-based tests for early detection
of colorectal cancer (CRC) [3], information on sex and age
of cases and controls was given in approximately half of the
studies. While many studies used convenience samples of
controls (such as patients with benign diseases or blood do-
nors) that were on average considerably younger than
cases, matching of controls by age or sex and age was ex-
plicitly reported in four studies [4e7].
However, for valid judgment of the specificity of cancer
early detection markers, the controls should be representa-
tive of cancer-free people from the screening population,
who might differ from the cases with respect to sex and
age. For example, because of the strong rise of both cancer
incidence and prevalence with age [8,9], cases would
almost always be expected to be on average older than non-
cases among screening participants. Likewise, the strong
 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208
What is new?
 Matching of controls to the sex and age distribu-
tion of cancer cases is commonly used, and the de-
gree of matching is often perceived to be a quality
criterion in studies evaluating cancer early detec-
tion tests.
 The authors show that, in contrast to widespread
belief, ‘‘perfect matching’’ may in fact lead to bi-
ased estimates of specificity, and they illustrate
the potential merits and pitfalls of matching using
the example of colorectal cancer screening studies.
sex differences in incidence and prevalence seen for many
forms of cancer, such as lung, stomach, and CRC or skin
melanoma, would be expected to result in major differences
in the sex distribution of cases and cancer-free controls
[8,9]. Unsurprisingly, age and sex were found to be strongly
associated with findings of cancer and precancerous lesions
in many cancer screening studies [10e12].
Implications, pros, and cons of matching as a tool to pre-
vent confounding or enhance precision and power have
been widely addressed in the context of assessing risk fac-
tor effects in cohort and caseecontrol studies, respectively
[13e15]. To our knowledge, no previous article has
addressed the implications of matching in studies aiming
to assess the performance characteristics of screening tests.
In this article, we aim to assess potential implications of
such matching. We illustrate by an empirical example from
the setting of CRC screening that matching of controls to
the sex and age distribution of cases might lead to biased
estimates of specificity if the sex and age distribution of
people in the screening population vary between those with
and without the disease and specificity likewise varies
according to sex and age. Our illustration is based on data
from the German national screening colonoscopy program
[16,17] and from a study evaluating two immunochemical
fecal occult blood tests (iFOBTs) among participants of this
program [18,19].
2. Methods
2.1. Databases
Data from the German national screening colonoscopy
registry were used to assess the sex and age distribution of
CRC cases and cancer-free participants of screening
colonoscopy. Colonoscopy is the current gold standard for
diagnosis of CRC. In Germany, screening colonoscopy is
offered as a primary screening examination for early
detection and prevention of CRC since October 2002.
Women and men are eligible for a first screening colono-
scopy from the age of 55 years. If this first screening
203
colonoscopy is conducted before 65 years of age, a second
screening examination will be offered 10 years later. Certifi-
cation to conduct screening colonoscopy is tightly regulated
on the basis of extensive previous training and experience,
and maintenance of certification is contingent on conducting
at least 200 colonoscopies and 10 polypectomies per year
that are subject to rigorous quality control. Histopathologic
examination is performed decentrally by certified pathology
laboratories.
Details on the national German screening colonoscopy
registry have previously been reported [16,17]. Briefly, the
results of all screening colonoscopies are reported on a stan-
dardized form by the physicians. Reporting is considered
virtually complete as it is a prerequisite for reimbursement
for colonoscopies by the health insurance funds. The re-
gistry includes only colonoscopies conducted as primary
screening examinations (i.e., colonoscopies conducted for
work-up of results from other tests, such as positive results
of fecal occult blood tests (FOBTs), because of symptoms
or for surveillance of previous findings, which are sepa-
rately reimbursed as ‘‘therapeutic colonoscopies’’ are not
included). Items reported include basic sociodemographic
variables and information on findings at colonoscopy. The
reporting forms are scanned, processed, and checked for
completeness and plausibility using standardized algorithms
at regional data centers before anonymized transfer to the
national data center is performed. Approximately 3% of el-
igible people participate in screening colonoscopy each
year, which translates to an expected participation rate of
25e30% during the 10-year time window foreseen for this
screening option. For this analysis, we used data from
1,007,814 first-time screening colonoscopies in 2006 and
2007. This time window was chosen as it corresponds to
the time window of data collection for the early detection
marker evaluation study described in the following.
iFOBTs are increasingly recommended and used for CRC
screening because of advantages in test performance and
acceptance over traditional guaiac-based FOBTs [20e24].
Estimates of the specificity of two iFOBTs were derived
from the BLITZ study, a study among participants of screen-
ing colonoscopy in Southern Germany which has been de-
scribed in detail elsewhere [18,19,25e27]. Briefly, 1,785
participants were recruited in 20 gastroenterology practices
between January 2006 and December 2007 according to
a protocol approved by the ethics committees of the Medical
Faculty Heidelberg of the University of Heidelberg and phy-
sicians’ chambers of Baden-W€urttemberg, Rheinland-Pfalz,
and Hessen. Patients were informed about the study at a
preparatory visit in the practice, typically about 1 week
before colonoscopy. They were asked to provide a stool sam-
ple before bowel preparation for colonoscopy, which was
used for the evaluation of multiple stool-based early detec-
tion markers, including multiple qualitative [18,25] and
quantitative iFOBTs [19,26,27] in a central laboratory.
Results for two quantitative tests (RIDASCREEN Haemo-
globin and RIDASCREEN Haemo-/Haptoglobin Complex;
204 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208

R-Biopharm AG, Darmstadt, Germany [28]) are used for il- 2. The selection of a sample of CRC-free participants
lustration here. The lower detection limit and the cut point with matching to the sex and age distribution of
for positivity given by the manufacturer are 0.42 and CRC cases (‘‘matched sampling’’).
2 mg/g stool for the hemoglobin test and 0.38 and 2 mg/g 3. ‘‘convenience sampling’’ of controls without specific
stool for the hemo-/haptoglobin complex test, respectively. attention to their sex and age distribution; here,
Furthermore, patients were asked to fill out a standardized a range of possible values of specificity was derived
questionnaire. Colonoscopy and histology reports were col- from the range of specificity estimates for the various
lected, and relevant data were extracted in a standardized subgroups defined by sex and age. In addition, we
manner. The latter was done independently by two trained provide an example of a specificity estimate expected
investigators who were blinded with respect to test results, from a relatively young convenience sample which
and the potential discrepancies were resolved by consensus. was derived by including only men and women aged
younger than 60 years in the control group.
2.2. Statistical analyses
The sex and age distribution (categories were 55e59, 3. Results
60e64, 65e69, 70e74, 75e79, and 80þ years) and mean
ages of screening colonoscopy participants with and The sex and age distribution of participants of the Ger-
without CRC were derived from the national registry by man screening colonoscopy program in 2006 and 2007
descriptive statistics. Specificities of the iFOBTs by sex with and without CRC is shown in Table 1. Forty-four per-
and age were derived from the BLITZ study. Among cent of cancer cases were aged 70 years or older compared
1,785 participants enrolled in BLITZ in 2006 and 2007, with 24% of participants without CRC. Conversely, only
the following exclusions were made to ensure conditions 12.7% of cancer patients were younger than 60 years of
of a screening setting and minimize potential misclassifi- age compared with 28.6% of screening participants with-
cation because of imperfect colonoscopy: visible rectal out CRC. These differences resulted in a mean age differ-
bleeding or previous positive FOBT result (n 5 111), in- ence of almost 4 years (68.6 vs. 64.7 years, respectively).
flammatory bowel disease (n 5 13), previous colonoscopy Also, the proportion of men was much higher (59.6%)
in the past 5 years (n 5 117), stool sampling after colono- among CRC cases than among participants without CRC
scopy (n 5 65), inadequate bowel preparation for colono- (45.7%).
scopy (n 5 79), and incomplete colonoscopy (n 5 22). In
addition, we excluded 48 patients with pseudopolyps or
Table 1. Age and sex distribution of participants of screening
histologically undefined polyps. After further exclusion
colonoscopy with and without colorectal cancer
of 1 participant with missing information on age, 6 parti-
Colorectal cancer
cipants with missing iFOBT results, and 11 cases with
CRC, 1,312 participants were retained for estimating Yes No
specificities by sex and age. To ensure reasonably precise Sex Age (yr) n % n %
specificity estimates, three rather than six age categories, Men 55e59 767 7.4 119,878 12.0
were used (!60, 60e69, and 70þ years). Multiple logis- 60e64 1,047 10.1 100,787 10.1
tic regression with test positivity as dependent variable 65e69 1,694 16.4 120,485 12.1
70e74 1,377 13.3 69,297 6.9
and sex and age as independent variables was used to test
75e79 862 8.3 33,203 3.3
for associations of both variables with specificity. Further- 80þ 404 3.9 11,481 1.2
more, specificities were also evaluated for a number of Total 6,151 59.6 455,131 45.6
higher cut points (6, 10, 14 mg/g stool) besides the one Mean (yr) 68.5 65.0
recommended by the manufacturer (2 mg/g stool) because Women 55e59 547 5.3 165,529 16.6
60e64 665 6.4 118,013 11.8
the latter yielded specificities for some subgroups that
65e69 1,071 10.4 132,926 13.3
would typically be regarded as too low for population- 70e74 871 8.4 73,760 7.4
based screening. 75e79 610 5.9 36,406 3.6
Finally, specificities were calculated for each cut point 80þ 409 4.0 15,725 1.6
that would be expected with the following sampling strate- Total 4,173 40.4 542,359 54.4
Mean (yr) 68.9 64.5
gies of controls:
Total 55e59 1,314 12.7 285,407 28.6
1. The complete inclusion of all CRC-free screening 60e64 1,712 16.6 218,800 21.9
65e69 2,765 26.8 253,411 25.4
colonoscopy participants meeting the criteria outlined
70e74 2,248 21.8 143,057 14.3
in the preceding paragraph or selection of a random 75e79 1,472 14.3 69,609 7.0
sample of these participants; we will refer to this 80þ 813 7.9 27,206 2.7
strategy as ‘‘correct sampling’’ as specificity is de- Total 10,324 100 997,490 100
rived from a sample that is expected to be representa- Mean (yr) 68.6 64.7
tive of the CRC-free screening population. German national screening colonoscopy registry, 2006e07.
 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208 205

Table 2. Specificity (%) of two immunologic fecal occult blood tests (RIDASCREEN Haemoglobin and RIDASCREEN Haemo-/Haptoglobin Complex)
by sex, age, and cut point of positivity
Test
Hemoglobin Hemo-/haptoglobin complex
Age (yr) Cut point (mg/g stool) Cut point (mg/g stool)
Sex Category Mean n 2 6 10 14 2 6 10 14
Men !60 55.8 215 82.3 92.6 91.5 96.3 86.5 94.9 96.7 96.7
60e69 64.6 316 81.3 86.7 87.7 90.2 83.2 92.4 95.9 97.8
70þ 73.6 130 79.9 80.0 83.3 86.9 81.5 90.0 92.3 92.3
Women !60 55.8 246 91.5 95.3 97.6 98.4 94.3 97.6 98.8 98.8
60e69 64.5 309 87.7 93.9 94.5 95.8 88.0 96.8 98.1 98.4
70þ 73.0 96 83.3 93.8 93.8 93.8 84.4 89.6 95.8 96.9
BLITZ study, Germany, 2006e07.

The specificity of the iFOBT by sex, age, and cut point
of positivity among participants of the BLITZ study is
shown in Table 2. Overall, 1,312 participants without
CRC with a mean age of 63.0 years were included, of
whom 50.4% were males. Sex- and age-specific estimates
of specificity are based on between 96 and 316 participants
per subgroup. As expected, specificity increased with
increasing cut points. For both iFOBTs and each cut point,
specificity was generally lower in men than that in women
and decreased with age between both sexes. Furthermore,
both sex and age were independent significant predictors
of specificity in multiple logistic regression models with
test result as dependent variable (P ! 0.05 in each case,
data not shown).
As a result, matching of controls to the sex and age dis-
tribution of cases (matched sampling) would be expected to
lead to underestimation of specificity for all cut points as-
sessed. For example, specificities of 83.0, 88.4, and
89.6% of the hemoglobin test would be expected at cut
points yielding true specificities of 84.8, 90.8, and 91.8%,
respectively (Table 3). This underestimation of specificity
in absolute terms by 1.8, 2.4, and 2.2 percent units would
correspond to a relative increase of the false-positive rate
by 12, 26, and 27%, respectively. Not caring about the
sex and age distribution in the selection of controls (conve-
nience sampling) could result in under- or overestimation
of specificity. A range of values of specificity that might
be obtained by convenience sampling are given by the
highest and lowest values of specificity observed in the
subgroups defined by sex and age. Ranges were generally
rather wide, and specificities at a given cut point varied
by up to 15.3 percent units depending on the control group
chosen. In the example of a relatively young convenience
sample of controls (mean age, 55.8 years) which was
obtained by restricting the analysis to men and women
younger than 60 years of age, specificity would have been
overestimated at each cut point by between 2.4 and 4.5 per-
cent units with the hemoglobin test and between 0.6 and 4.1
percent units with the hemo-/haptoglobin complex test.
4. Discussion
Our empirical example illustrates that matching of con-
trols to the sex and age distribution of cases might lead to

Table 3. Expected specificity (%) of immunologic fecal occult blood test (RIDASCREEN Hemoglobin and RIDASCREEN Haemo-/Haptoglobin
Complex) according to the sampling of controls and cut points of positivity
Cut point (mg/g stool)
Test Sampling 2 6 10 14
a
Hemoglobin Correct sampling 84.8 90.8 91.8 93.8
Matched samplingb 83.0 88.4 89.6 91.8
Convenience sampling
Rangec 79.9e91.5 80.0e95.3 83.3e97.6 86.9e98.4
Exampled 87.2 94.4 96.3 97.4
Hemo-/haptoglobin complex Correct samplinga 86.6 94.0 96.6 97.2
Matched samplingb 84.6 92.5 95.5 96.9
Convenience sampling
Rangec 81.5e94.3 89.6e97.6 92.3e98.8 92.3e99.0
Exampled 90.7 96.3 97.8 97.8
German national screening colonoscopy registry and BLITZ study, Germany, 2006e07.
a
Sex and age distribution corresponds to that in carcinoma-free screening participants.
b
Matching to sex and age distribution of colorectal cancer cases.
c
Range of values observed in subgroups defined by sex and age (Table 2).
d
Example of a relatively young convenience sample (mean age, 55.8 years) that was derived by restricting the analysis to men and women
younger than 60 years of age.
206 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208
biased estimates of the specificity of screening tests if the sex
and age distribution of cases and controls in the screening
population vary and specificity likewise varies according to
sex and age. The former condition is expected to commonly
hold and has repeatedly been demonstrated in cancer screen-
ing because of the strong increase in cancer incidence and
prevalence with age for almost all cancers and common var-
iation in cancer incidence and prevalence between men and
women [8e12]. The latter condition may also frequently
apply and was quite pronounced for the iFOBTs used for il-
lustration in our analysis. Other common examples include
measurements of pepsinogen and inflammatory markers in
peripheral blood, which have been suggested as screening
markers for gastric and a variety of other cancers [29e31]
and which are well known to show a major variation by
sex and age [32,33]. Therefore, the sex and age distribution
of controls should reflect that of cancer-free screening partic-
ipants rather than that of cancer cases.
Ideally, the evaluation of cancer screening tests should
be done in screening populations with direct sampling of
cases and controls from the subpopulations with and with-
out cancer. The German screening colonoscopy program
provides such a setting for the evaluation of early detection
markers of CRC because subpopulations with and without
CRC in the screening population can be reliably distin-
guished by colonoscopy, which is considered the diagnostic
gold standard in this context. Most studies aiming to eval-
uate cancer screening tests are conducted in different set-
tings, however, starting with the identification and
sampling of cancer cases for which adequate controls are
then sought. Often, cases are identified in clinical settings
(e.g., after the diagnosis and before the start of therapy),
and convenience samples are used as controls, such as
patients undergoing similar diagnostic procedures but
found free of the cancer of interest or other healthy volun-
teers, such as blood donors. Such convenience sampling
may sometimes lead to very large discrepancies in sex
and age distribution of cases and controls, far beyond those
justified by true differences in the screening population.
For example, in the aforementioned review of studies on
blood-based tests for CRC early detection [3], the majority
of studies used healthy volunteers, patients with benign
diseases, or a combination of both as controls. Where re-
ported, the mean age of controls was mostly much lower
than that of CRC cases, with differences exceeding 20 years
in several studies [34e38]. In such situations, matching of
controls to the sex and age distribution of cases might have
provided less biased estimates of specificity. Nevertheless,
direct sampling from the cancer-free screening population,
which would avoid additional threats of validity that may
result from convenience sampling (a detailed discussion
of which is beyond the scope of this article) and which
may typically yield some difference in sex and age distribu-
tion of cases and controls, should be the preferred strategy
to be applied whenever possible.
It should be noted, however, that the arguments outlined
in this article refer to studies aiming to describe the sensi-
tivity and specificity of diagnostic tests in the screening set-
ting. In other contexts, preferences may be different. In
particular, there might be studies primarily aiming to assess
to what extent a test, whose results vary by age and sex,
might have any independent diagnostic value (beyond diag-
nostic value mediated by its association with age and sex)
to distinguish people with and without cancer. In such stud-
ies, matching for age and sex might be a method of choice
indeed. Vice versa, the use of convenience samples of con-
trols whose age and sex distribution strongly vary from that
of cases might inappropriately attribute diagnostic value to
tests whose results vary by age and sex independent of the
presence of cancer.
Regarding the specific context of our empirical example,
a few additional issues require further discussion. In many
studies on early detection markers of CRC, cases are re-
cruited in the clinical setting rather than in a screening set-
ting, and the age distribution of those cases is likely to be
even further shifted toward higher ages compared with
CRC-free controls from the screening population. Accord-
ing to data from population-based cancer registries from
Germany [39], the estimated proportion of patients diag-
nosed with CRC in Germany in 2006 and 2007 who were
70 years or older at the time of diagnosis was 57.6%, that
is, substantially higher than the corresponding proportion
of 44% observed among CRC patients identified by screen-
ing colonoscopy (Table 1). On the other hand, cancer-free
participants of screening colonoscopy in our empirical
example may be older on average than the typical cancer-
free CRC screening population because screening colono-
scopy is offered from the age of 55 years only in Germany,
whereas screening is recommended from the age of 50
years by expert panels [40e42], and screening by FOBT
is offered starting from the age of 50 years in Germany.
Therefore, the age gap between CRC cases and cancer-
free controls may even be larger, and the potential bias
from matching by age may be more severe in other settings
than in the one assessed in our empirical example. Further
limitations to generalizability might arise from potential
differences in self-selection of participants of screening
colonoscopy and users of other noninvasive options of pri-
mary screening tests [43,44].
In our example, specificity was defined by the propor-
tion of negative tests among those without CRC. Ideally,
however, screening tests for CRC should also detect ad-
vanced adenomas, the precursors of CRC, and possibly
even other (nonadvanced) adenomas. It might therefore
be argued that specificity should be determined among
those free of advanced neoplasms (or even free of any neo-
plasms) only. Repeating our analyses with these alternative
definitions of specificity yielded substantially higher levels
of specificity at all cut points assessed, along with an even
slightly larger gap in mean age between CRC cases and
 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208
controls and a similar bias by matching of controls to the
sex and age distribution of cases.
Specificity decreased with age and was lower among
men than among women for the specific tests evaluated
in our example, leading to underestimation of specificity
by matched sampling and overestimation of specificity in
case of a relatively young convenience sample. The de-
crease in specificity with age might be explained by an in-
creasing risk of bleeding from other gastrointestinal lesions
with older age. Likewise, lower specificity among men than
among women might result from higher prevalences of ul-
cer or other gastrointestinal bleeding sources and higher
prevalences of aspirin use among men than among women
[45]. Similar variation of specificity by sex and age would
be expected for other FOBTs. For other types of tests, inde-
pendence of specificity from sex and age or even reverse
patterns might also be conceivable. If sex and age are unre-
lated to specificity, then matching by these factors does not
introduce bias, but it is also unnecessary as it then does not
have any impact on specificity estimates.
In our study, performance of colonoscopy, which is con-
sidered the gold standard for the detection of CRC, allowed
the distinction of screening participants with and without
CRC at high reliability. In many other screening settings,
delineation of the cancer-free subpopulation may not be
as straightforward or not be possible at all. In such settings,
the age and sex distribution of the entire screening popula-
tion may often be a reasonable proxy for the age and sex
distribution of the cancer-free screening population, given
the low prevalence of cancer in most screening settings.
Our study has specific strengths and limitations.
Strengths include the very large database of the German na-
tional screening colonoscopy registry used to assess sex and
age distribution of CRC cases and CRC-free controls and
the evaluation of the tests in a true screening setting. We
presented a detailed empirical illustration of variation of
specificity by sex and age for two tests only. However, very
similar patterns were seen for both tests and in fact would
be expected for FOBTs in general, the so far best estab-
lished noninvasive tests for CRC screening. In most previ-
ous evaluations of other CRC early detection markers, no
stratification by sex and age was done, and for many of
those markers, it is unknown to what extent their specificity
may vary by sex and age.
Despite its limitations, our article illustrates the poten-
tial merits and pitfalls of matching of controls in the eval-
uation of cancer screening tests. Although matching of
controls to the sex and age distribution of cases may help
to avoid potentially even larger bias by the use of conve-
nience samples whose sex and age distribution may be
even further away from that of cancer-free people from
the screening population, such matching may still result
in biased estimates of specificity. Along the same lines,
agreement of sex and age distribution of cases and controls
should not be regarded as a quality criterion of studies aim-
ing to assess the sensitivity and specificity of tests in cancer
207
screening settings. Although extreme differences in such
distribution may be indicative of the use of inappropriate
convenience samples, full agreement enforced by matching
is also often undesirable as it typically hinders controls
from being representative of the cancer-free screening
population.
Ideally, controls should be a random sample of the
cancer-free screening population. In studies in which ran-
dom sampling of controls from the cancer-free study popu-
lation is not possible, matching of controls to the age and
sex distribution of the cancer-free screening population
(which can typically be closely approximated by the age
and sex distribution of the total screening population) rather
than matching to the age and sex distribution of cases might
be the method of choice. In studies in which the age and sex
distribution of controls differs from that of the cancer-free
screening participants, adjustment to the age and sex distri-
bution of the cancer-free screening population by appropri-
ate weighting of age- and sex-specific estimates of
specificity might be considered to prevent the type of pos-
sible bias outlined in this article.
Acknowledgments
The authors acknowledge excellent contributions in the
conduction of the BLITZ study by Isabel Lerch, Sabrina
Hundt, Ulrike Haug, and the cooperating gastroenterology
practices. The authors are grateful to Labor Limbach
(Heidelberg) for laboratory analyses of the iFOBT.
References
[1] Bosch LJ, Carvalho B, Fijneman RJ, Jimenez CR, Pinedo HM, van
Engeland M, et al. Molecular tests for colorectal cancer screening.
Clin Colorectal Cancer 2011;10:8e23.
[2] Luo X, Burwinkel B, Tao S, Brenner H. MicroRNA signaturesd
novel biomarker for colorectal cancer? Cancer Epidemiol Biomark
Prev 2011;20:1272e86.
[3] Tao S, Hundt S, Haug U, Brenner H. Sensitivity estimates of blood
based tests for colorectal cancer detection: impact of overrepresenta-
tion of advanced stage disease. Am J Gastroenterol 2011;106:242e53.
[4] Dudouet B, Jacob L, Beuzeboc P, Magdalenat H, Robine S,
Chapuis Y, et al. Presence of villin, a tissue-specific cytoskeletal pro-
tein, in sera of patients and an initial clinical evaluation of its value
for the diagnosis and follow-up of colorectal cancers. Cancer Res
1990;50:438e43.
[5] Schiedeck TH, Wellm C, Roblick UJ, Broll R, Bruch HP. Diagnosis
and monitoring of colorectal cancer by L6 blood serum polymerase
chain reaction is superior to carcinoembryonic antigen-enzyme-
linked immunosorbent assay. Dis Colon Rectum 2003;46:818e25.
[6] Holten-Andersen MN, Christensen IJ, Nielsen HJ, Lilja H,
Murphy G, Jensen V, et al. Measurement of the noncomplexed free
fraction of tissue inhibitor of metalloproteinases 1 in plasma by im-
munoassay. Clin Chem 2002;48:1305e13.
[7] Leung WK, To KF, Man EP, Chan MW, Bai AH, Hui AJ, et al. Quan-
titative detection of promoter hypermethylation in multiple genes in
the serum of patients with colorectal cancer. Am J Gastroenterol
2005;100:2274e9.
[8] Lutz JM, Francisci S, Mugno E, Usel M, Pompe-Kirn V,
Coebergh J-W, et al. Cancer prevalence in Central Europe: the
EUROPREVAL Study. Ann Oncol 2003;14:313e22.
208 H. Brenner et al. / Journal of Clinical Epidemiology 66 (2013) 202e208
[9] Forman D, Stockton D, Møller H, Quinn M, Babb P, De Angelis R,
et al. Cancer prevalence in the UK: results from the EUROPREVAL
Study. Ann Oncol 2003;14:648e54.
[10] Segnan N, Senore C, Andreoni B, Aste H, Bonelli L, Crosta C, et al.
Baseline findings of the Italian multicenter randomized controlled
trial of ‘‘once only sigmoidoscopy’’dSCORE. J Natl Cancer Inst
2002;94:1763e72.
[11] Hocking WG, Hu P, Oken MM, Winslow SD, Kvale PA, Prorok PC,
et al. Lung cancer screening in the randomized Prostate, Lung, Colo-
rectal, and Ovarian (PLCO) Cancer Screening Trial. J Natl Cancer
Inst 2010;102:722e31.
[12] Maisonneuve P, Bagnardi V, Bellomi M, Spaggiari L, Pelosi G,
Rampinelli C, et al. Lung cancer risk prediction to select smokers
for screening CTda model based on the Italian COSMOS trial.
Cancer Prev Res 2011;4:1778e89.
[13] Kupper LL, Karon JM, Kleinbaum DG, Morgenstern H, Lewis DK.
Matching in epidemiologic studies: validity and efficiency consider-
ations. Biometrics 1981;37:271e91.
[14] Costanza MC. Matching. Prev Med 1995;24:425e33.
[15] St€urmer T, Brenner H. Degree of matching and gain in power and
efficiency in case-control studies. Epidemiology 2001;12:101e8.
[16] Brenner H, Hoffmeister M, Stegmaier C, Brenner G, Altenhofen L,
Haug U. Risk of progression of advanced adenomas to colorectal can-
cer by age and sex: estimates based on 840,149 screening colonos-
copies. Gut 2007;56:1585e9.
[17] Brenner H, Altenhofen L, Hoffmeister M. Sex, age and birth cohort
effects in colorectal neoplasms: a cohort analysis. Ann Intern Med
2010;152:697e703.
[18] Hundt S, Haug U, Brenner H. Comparative evaluation of immuno-
chemical fecal occult blood tests for colorectal adenoma detection.
Ann Intern Med 2009;150:162e9.
[19] Brenner H, Tao S, Haug U. Low-dose aspirin use and performance of
immunochemical fecal occult blood tests. JAMA 2010;304:2513e20.
[20] Guittet L, Bouvier V, Mariotte N, Vallee JP, Arsene D, Boutreux S,
et al. Comparison of a guaiac based and an immunochemical faecal
occult blood test in screening for colorectal cancer in a general aver-
age risk population. Gut 2007;56:210e4.
[21] van Rossum LG, van Rijn AF, Laheij RJ, van Oijen MG, Fockens P,
van Krieken HH, et al. Random comparison of guaiac and immuno-
chemical fecal occult blood tests for colorectal cancer in a screening
population. Gastroenterology 2008;135:82e90.
[22] Park DI, Ryu S, Kim YH, Lee SH, Lee CK, Eun CS, et al. Comparison
of guaiac-based and quantitative immunochemical fecal occult blood
testing in a population at average risk undergoing colorectal cancer
screening. Am J Gastroenterol 2010;105:2017e25.
[23] van Dam L, Kuipers EJ, van Leerdam ME. Performance improve-
ments of stool-based screening tests. Best Pract Clin Gastroenterol
2010;24:479e92.
[24] Duffy MJ, van Rossum LG, van Turenhout ST, Malminiemi O,
Sturgeon C, Lamerz R, et al. Use of faecal markers in screening
for colorectal neoplasia: a European group on tumor markers position
paper. Int J Cancer 2011;128:3e11.
[25] Brenner H, Haug U, Hundt S. Inter-test agreement and quantitative
cross-validation of immunochromatographical fecal occult blood
tests. Int J Cancer 2010;127:1643e9.
[26] Brenner H, Haug U, Hundt S. Sex differences in performance of fecal
occult blood testing. Am J Gastroenterol 2010;105:2457e64.
[27] Haug U, Hundt S, Brenner H. Quantitative immunochemical fecal oc-
cult blood testing for colorectal adenoma detection: evaluation in the
target population of screening and comparison with qualitative tests.
Am J Gastroenterol 2010;105:682e90.
[28] http://www.r-biopharm.com/product_site.php?product_range5Clinical
Diagnostics&product_class_one5R2FzdHJvZW50ZXJvbG9neQ55
&product_class_two5Q29sb24gQ2FuY2VyIFByZXZlbnRpb245&,
last accessed 3 January 2012.
[29] Shiotani A, Iishi H, Uedo N, Kumamoto M, Nakae Y, Ishiguro S,
et al. Histologic and serum risk markers for noncardia early gastric
cancer. Int J Cancer 2005;115:463e9.
[30] Mukoubayashi C, Yanaoka K, Ohata H, Arii K, Tamai H, Oka M,
et al. Serum pepsinogen and gastric cancer screening. Intern Med
2007;46:261e6.
[31] Tao S, Haug U, Kuhn K, Brenner H. Comparison and combination of
blood-based inflammatory markers with faecal occult blood tests
for non-invasive colorectal cancer screening. Br J Cancer 2012;
106:1424e30.
[32] Sun LP, Gong YH, Wang L, Yuan Y. Serum pepsinogen levels and
their influencing factors: a population-based study in 6990 Chinese
from North China. World J Gastroenterol 2007;13:6562e7.
[33] Singh T, Newman AB. Inflammatory markers in population studies of
aging. Ageing Res Rev 2011;10:319e29.
[34] van Kamp GJ, von Mensdorff-Pouilly S, Kenemans P, Verstraeten R,
Yedema CA, Wobbes T, et al. Evaluation of colorectal
cancer-associated mucin CA M43 assay in serum. Clin Chem 1993;
39:1029e32.
[35] Huber K, Kirchheimer JC, Sedlmayer A, Bell C, Ermler D,
Binder BR. Clinical value of determination of urokinase-type plas-
minogen activator antigen in plasma for detection of colorectal
cancer: comparison with circulating tumor-associated antigens
CA 19-9 and carcinoembryonic antigen. Cancer Res 1993;53:
1788e93.
[36] Hyodo I, Doi T, Endo H, Hosokawa Y, Nishikawa Y, Tanimizu M,
et al. Clinical significance of plasma vascular endothelial growth fac-
tor in gastrointestinal cancer. Eur J Cancer 1998;34:2041e5.
[37] Douard R, Le Maire V, Wind P, Sales JP, Dumas F, Fayemendi L,
et al. Carcinoembryonic gene member 2 mRNA expression as
a marker to detect circulating enterocytes in the blood of colorectal
cancer patients. Surgery 2001;129:587e94.
[38] Douard R, Moutereau S, Serru V, Sales JP, Wind P, Cugnenc PH,
et al. Immunobead multiplex RT-PCR detection of carcinoembryonic
genes expressing cells in the blood of colorectal cancer patients. Clin
Chem Lab Med 2005;43:127e32.
[39] Gesellschaft der epidemiologischen Krebsregister in Deutschland
e.V. (GEKID). GEKID-Atlas. http://www.gekid.de/, last accessed
2 January 2012.
[40] U.S. Preventive Services Task Force. Screening for colorectal cancer:
U.S. preventive services task force recommendation statement. Ann
Intern Med 2008;149:627e37.
[41] Levin B, Lieberman DA, McFarland B, Andrews KS, Brooks D,
Bond J, et al. Screening and surveillance for the early detection of
colorectal cancer and adenomatous polyps, 2008: a joint guideline
from the American Cancer Society, the US Multi-Society Task Force
on Colorectal Cancer, and the American College of Radiology. Gas-
troenterology 2008;134:1570e95.
[42] Schmiegel W, Reinacher-Schick A, Arnold D, Graeven U,
Heinemann V, Porschen R, et al. [Update S3 guidelines colorectal
cancer 2008]. in German. Z Gastroenterol 2008;46:799e840.
[43] Powell AA, Burgess DJ, Vernon SW, Griffin JM, Grill JP,
Noorbaloochi S, et al. Colorectal cancer screening mode preferences
among US veterans. Prev Med 2009;49:442e8.
[44] Stock C, Brenner H. Utilization of lower gastrointestinal endoscopy
and fecal occult blood test in 11 European countries: evidence from
the Survey of Health, Aging and Retirement in Europe (SHARE).
Endoscopy 2010;42:546e56.
[45] Garcia Rodriguez LA, Hernandez-Diaz S. Risk of uncomplicated
peptic ulcer among users of aspirin and nonaspirin nonsteroidal
anti-inflammatory drugs. Am J Epidemiol 2004;159:23e31.