Report in English with a Dutch summary (KCE reports 45A)

92 Screening for Colorectal Cancer KCE reports vol.45
On the other hand, there is the issue of overinvestigation in the group with
false-positive tests. Despite the fact that the guaiac tests are very insensitive for
upper gastrointestinal bleeding, there remains some concern.
5.6.2 False-negative results
In addition, false-negative results caused by the low sensitivity of the FOBT and
the propensity of sigmoidoscopy to miss proximal cancers might falsely reassure
individuals (i.e. the "certificate of health effect") and lead to delayed cancer
diagnosis and poorer outcome.
Some patients present with fully developed cancers within 1 - 4 yr of a
colonoscopy that apparently cleared the colon of neoplasia. These events may
result in medical-legal action against gastroenterologists, generally based on an
assumption of negligent technical performance of the procedure. Alternative
explanations for the development of interval cancers include variable growth
rates of colorectal cancers, the inherent miss rate of the procedure, even when
optimal examination techniques are used, and the possibility of flat lesions that
are not readily detected by standard colonoscopic techniques. Issues relevant to
reduction of medical-legal risks associated with interval cancers after clearing
colonoscopy include informed consent, documentation of cecal intubation,
appropriate description of preparation, documentation of examination time and
technique, and attention to potential atypical neoplasms 292.
5.6.3 Studies on CRC screening harms
The Nottingham group has addressed these forgoing issues by examining the
investigation and treatment-related mortality and the stage at presentation of
the interval cancers 296. There were no colonoscopy-related deaths and five
deaths after surgery for screen-detected cancers; this represents a 2% operative
mortality at a time when mortality after elective colorectal cancer surgery in
the United Kingdom was estimated to be around 5% by a large national audit 395.
Furthermore, the stage distribution of the interval cancers (cancers that were
diagnosed after a negative FOBT or colonoscopy) was similar to that of the
cancers in the control group, but the survival was significantly better than that
for the control cancers. Nevertheless, these concerns have been highlighted by
the finding that all-cause mortality is not affected by colorectal cancer screening
and indeed, in the Nottingham study it was found to be even increased in the
group offered screening 15. However, colorectal cancer only accounts for around
2% of all deaths, and a 15% reduction in disease-specific death rate could only
be expected to reduce all-cause mortality by 0,3%. To demonstrate a difference
of this size with statistical power would require a trial too big to be feasible.
Furthermore, unlike the difference in disease-specific mortality, the excess of
all-cause deaths observed in the group offered screening was not statistically
significant and therefore likely to represent a chance finding.
To try to rationalize the fear that ignoring a positive FOBT in the face of a
normal colonoscopy might be seen as negligent, if significant upper
gastrointestinal pathology is missed, the Nottingham group looked at a cohort
of 283 FOBT positive cases without neoplastic disease diagnosed at
colonoscopy 396. Fourteen (5%) of these underwent upper gastrointestinal
endoscopy because of symptoms, and one was found to have gastric carcinoma.
The rest, who were asymptomatic, were followed up for a median period of five
years and only one, who had persistent symptoms after a previous partial
gastrectomy, was subsequently diagnosed as having gastric cancer. Thus, the
evidence supports a strategy of reassuring the majority of those who have a