Supplementary Table 1: Risk of bias assessment

First author, year, Ref. / RPC / DLR / DAA / RRP / DCC / PSC / IET / SUB / WDT / %ITS / %BP / %CC / ADP / SD
Morikawa 2005 (12) / - / -7 / - / + / + / + / + / - / ? / NA / ? / 98.2 / ++ / WHO
Nakazato 2006 (13) / - / -8 / - / + / + / ? / + / ? / ? / NA / ? / 98.6 / + / JSCCR
Graser 2009 (14) / +1 / NA / + / - / + / + / ? / + / ? / ? / ? / 99.4 / + / ESGAR+WGVC
Park 2010 (15) / +2 / - / + / + / + / + / + / ? / ? / 82 / ? / 99.9 / + / ?
Haug 2011 (16) / +3 / + / + / + / + / + / + / ? / ? / ? / ? / 98.7 / + / ?
Khalid-de Bakker 2011 (17) / +4 / + / + / + / + / + / ? / ? / + / 41 / 52 / 91.7 / + / ?
de Wijkerslooth 2012 (18) / - / + / + / + / + / + / + / ? / +9 / 22 / ? / 98.6 / +(+) / ?
Chiu 2013 (19) / +5 / + / + / + / + / + / + / ? / ? / NA / ? / 99.0 / + / WHO
Imperiale 2014 (20) / - / + / + / + / + / + / + / ? / ? / ? / ? / 99.0 / + / ?
Wong 2015 (21) / +6 / + / + / - / + / + / + / ? / + / ? / 13 / ? / ? / ?
Sum / 6 / 6 / 8 / 8 / 10 / 9 / 8 / 1 / 3 / / / / / / / /

Abbreviations:

RPC =reporting of potentially confounding variables

DLR =study reported a clear and common definition of the left and/or right colon. Definitions were rated as incorrect if the transverse colon (or segments proximal to it) was not assigned to the right colon or if the descending colon (or segments distal to it) was not assigned to the left colon.

DAA = study using the common definition of advanced adenomas (adenomas ≥1cm and/or high-grade dysplasia and/or (tubulo-)villous components

RRP = clear reporting of the study recruitment process with numbers and reasons of exclusions, ideally in a flow-chart

DCC = clear description of data collection and conduction of FIT

PSC = prespecified cutoff of quantitative FIT. (Qualitative FIT reporting results based on the cutoff recommended by the manufacturer or using the lower detection limit were rated as prespecified.)

IET = independent examination of the tests (laboratory staff members who evaluated FIT were blinded to colonoscopic findings or endoscopists were blinded from the results of FIT, whichever test was conducted first)

SUB = segmental unblinding used to avoid an overestimation of FS sensitivity

WDT = withdrawal time of the endoscope of at least 6 minutes

%ITS = percentage of considered participants that actually underwent colonoscopy

%BP = percentage of included participants with optimal bowel preparation

%CC = percentage of completed colonoscopy (cecal intubation rate)

ADP = assessment of diagnosis by (histo-)pathology; “+”, histology/at least one pathologist, “+(+)”, 2nd pathologist for 10% of the samples, “++”, at least 2 pathologists

SD = standard of diagnosis. Abbreviations: WHO, World Health Organization; JSCCR, Japanese Society for Cancer of the Colon and Rectum; ESGAR, European Society of Gastrointestinal and Abdominal Radiology; WGVC, Working Group on Virtual Colonoscopy.

“+”=fulfilled, “-”=not fulfilled, “?”=unclear, “/”, not applicable

1Age and sex of every participant together with findings on colonoscopy and FIT result reported

2Test results reported stratified by histology (high- and low-grade dysplasia), size and number of adenomas, and cancer stages

3Mean age and proportions of men, current NSAID users, subjects with neoplasms ≥1 cm, ≥1 neoplasm, ≥1 advanced neoplasm, pedunculated adenoma(s), and CRC reported

4Test results reported in relation to location of adenomas, gender, and age

5Results stratified by invasive cancer, carcinoma in situ and T1 cancer, and T2-T4 cancer

6Binary logistic regression models conducted including age, NSAID use, smoking

7Although a definition of the left and right colon was provided, it remained unclear which side neoplasms in the splenic flexure were assigned to.

8Nakazato et al. was the only publication that also assigned the descending colon to the right (proximal) side of the colon

9Withdrawal time>6 min in 99.8% of all colonoscopies

Supplementary Table 2: Alternative specificities of FIT alone and combined with FS for CRC

FIT alone or
no colo referral
Specificity for CRC(95% CI)
Morikawa 2005 (12) / 93 (93-94)
Nakazato 2006 (13) / 87 (86-88)
Graser 2009 (14) / 85 (80-88)
Park 2010 (15) / 90 (88-92)
Haug 2011 (16) / –
Khalid-de Bakker 2011 (17) / –
de Wijkerslooth 2012 (18) / 95 (93-96)
Chiu 2013 (19) / 94 (93-94)
Imperiale 2014 (20) / 93 (93-94)
Wong 2015 (21) / 91 (90-91)
Summary estimate / 92 (89-94)
Excluding Nakazato / 92 (90-94)

Advanced adenoma cases were interpreted as false-positives for the calculations of specificity for CRC.

Supplementary Figure. Assessment of publication bias using the method of Deeks et al. (43)

Abbreviations: CRC, colorectal cancer; AA, advanced adenoma; AN, advanced neoplasm (CRC or AA); DOR, diagnostic odds ratio [(sensitivity x specificity) / ((1-sensitivity) x (1-specificity))]; ESS, effective sample size [(4 x nnondiseased x ndiseased)/(nnondiseased + ndiseased)]. Note: One study on CRC(14) was excluded in the plot because it reported only one CRC and the DOR could not be computed.