Original article
12778
Colonoscopy screening compliance and outcomes in patients
with Lynch syndrome
K. Newton*, K. Green†, F. Lalloo†, D. G. Evans† and J. Hill*
*Department of General Surgery, Manchester Royal Infirmary, Manchester, UK and †Manchester Centre for Genomic Medicine, Central Manchester
University Hospitals NHS Foundation Trust, Saint Mary’s Hospital, Manchester, UK
Received 20 April 2014; accepted 15 July 2014; Accepted Article online 11 September 2014
Abstract
Aim Colonic surveillance reduces the lifetime risk of
colorectal cancer in patients with Lynch syndrome
(hereditary nonpolyposis colorectal cancer) from 60 to
80% to 10% and confers a 7-year survival advantage.
The British Society of Gastroenterologists recommends
colonoscopy at least every 2 years from the age of 25.
Currently in the UK, genetic diagnosis is made by a
regional genetics service, and screening recommenda-
tions are made to the referring clinician. The aim of this
study was to investigate compliance with and the effec-
tiveness of large bowel surveillance in Lynch syndrome.
Method A retrospective longitudinal study of Lynch
syndrome mutation carriers on the Regional Familial
Colorectal Cancer Registry under and not under screen-
ing was conducted. To investigate screening compli-
ance, patients were included if they were alive at the
start of the study. Data were gathered on timeliness,
quality and outcome of screening. To examine the
effectiveness of screening, the cumulative incidence of
colorectal cancer was estimated using Kaplan–Meier
curves and the screened population compared with
patients not being screened.
Results A total of 227 Lynch syndrome mutation carri-
ers were under screening at 26 hospitals. We assessed
Introduction
Lynch syndrome (hereditary nonpolyposis colorectal
cancer) is a cancer-predisposition syndrome inherited in
an autosomal dominant fashion. It is caused by a muta-
tion in one of the mismatch repair genes (MLH1, mutL
homologue; MSH2, mutS homologue 2; MSH6, mutS
Correspondence to: K. Newton, c/o Clinical Genetics Service, Manchester
Centre for Genomic Medicine, St Mary’s Hospital, Central Manchester
University Hospitals NHS Foundation Trust, Oxford Road, Manchester M13
9LW, UK.
E-mail: katynewton2012@doctors.org.uk
38 Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
doi:10.1111/codi.
439 colonoscopies for timeliness, of which 68% were
compliant (interval < 27 months). Compliance on the 1
November 2011 was 87%. The cumulative incidence of
colorectal cancer to the age of 70 was 25% (95% CI
17–32%) in the surveillance population and 81% (95%
CI 78–84%) in 689 mutation-positive patients not
being screened (P < 0.0001).
Conclusion Overall, 68% of colonoscopies were on
time. The incidence of colorectal cancer was greatly
reduced by screening but remained significant. Patients
with Lynch syndrome need proactive surveillance man-
agement.
Keywords Lynch syndrome, HNPCC, surveillance,
screening, colorectal cancer
What does this paper add to the literature?
This is the first UK study of compliance and cancer risk
in mutation-positive Lynch Syndrome. Compliance with
2-yearly surveillance is 67%. This surveillance pro-
gramme reduces the incidence of colorectal cancer
(CRC) from 81% to 25%. CRC incidence remains sig-
nificant and justifies screening managed to the same
standards as the National Bowel Cancer Screening Pro-
gramme.
homologue 6; PSM2, postmeiotic segregation increased
2) [1]. Lynch syndrome causes up to 4% of all colorec-
tal cancers (CRCs) and is the most prevalent form of
inherited CRC [2,3]. The condition has variable pene-
trance but has a cumulative lifetime risk of CRC of
60–80% if patients are not screened. Presymptomatic
genetic assessment and gene testing allows for the iden-
tification and surveillance of mutation carriers. Regular
colonoscopy provides the opportunity to prevent CRC
through the endoscopic removal of adenomas. When
cancers do occur, surveillance detects them at an earlier
and hence a more treatable stage.
K. Newton et al.
Current UK guidelines for large bowel surveillance
in Lynch syndrome patients recommend at least biennial
colonoscopy from the age of 25 [4]. Surveillance at 18-
month intervals may be more appropriate following
reports of interval cancers [5]. Retrospective descriptive
case–control series have shown that regular screening
reduces CRC incidence [2,6–8], tumour stage [6,9]
and CRC-specific [2,9–11] and overall mortality [2,12].
The only prospective controlled (nonrandomized) study
found that 3-yearly surveillance reduced CRC incidence
by 62% and improved survival by 65% in mutation-posi-
tive individuals [2]. A relative risk of CRC of 0.44 (95%
CI 0.215–0.9, P = 0.02) in 44 screened Lynch syn-
drome mutation carriers was reported compared with
46 unscreened mutation carriers. There were no CRC-
related deaths in the screened group, and the eight can-
cers diagnosed in this group were of a more favourable
stage than those diagnosed in the control group. The
5-year survival was 100% in the screened group and
54% in those who were not screened. Relative risk of
death in screened mutation-positive individuals com-
pared with controls was 0.35 (95% CI 0.122–0.999)
[2].
Current UK practice for individuals with suspected
familial CRC is for referral to a regional genetic medi-
cine service. Following diagnosis, surveillance recom-
mendations are made to the patient’s clinician.
Compliance with these recommendations in a UK pop-
ulation is unknown. Compliance with surveillance has
been reported in Europe and the United States and var-
ies between 63 and 93% in HNPCC family members
[13–19]. Several studies were questionnaire based
[15,20,21] thus required self-reporting of screening
attendance. Significant bias towards better compliance is
therefore likely. Many of these studies investigate sur-
veillance compliance in terms of at least one colonos-
copy attendance or attendance in the last 1–2 years
[15,17,18]. In Lynch syndrome the risk of CRC is life-
long. Patients must adhere to multiple colonoscopies
over a sustained period of time in order for the cancer
prevention to be entirely effective.
This aim of this study was: (i) to investigate the
objective long-term compliance with large bowel sur-
veillance in a large, well-defined, mutation-positive
Lynch syndrome population, and (ii) to examine the
effectiveness of screening by comparing the incidence of
CRC in patients being screened and those not being
screened.
Method
This is a retrospective longitudinal study of Lynch syn-
drome mutation-positive individuals listed on the Man-
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
HNPCC colonoscopy screening compliance
chester Familial Colorectal Cancer Registry. The
registry was established in 1999. All individuals referred
to the regional clinical genetics service with a family his-
tory of colorectal cancer are entered onto the registry
database. Relatives within the pedigree are also listed.
Results of tumour mismatch repair immunohistochemis-
try, microsatellite instability, mismatch repair germline
gene mutation testing, cancer incidence and pathology
are prospectively recorded. Dates of cancer diagnosis
and dates of last follow-up or death are prospectively
verified using the North West Cancer Intelligence Ser-
vice (NWCIS), family genetic records and the NHS
summary care records. The clinical genetics service
serves a population of six million and receives referrals
from GPs, colorectal surgeons, gastroenterologists and
oncologists. All individuals with suspected or confirmed
Lynch syndrome are offered biannual colonic surveil-
lance according to the British Society of Gastroenterol-
ogists guidance [4] either at the unit attached to the
regional genetics service or at their local hospital. Data
on surveillance are gathered prospectively at biannual
registry review (telephone interview by a genetic coun-
sellor).
To investigate current compliance with surveillance
all pathogenic Lynch syndrome gene mutation carriers
and obligate carriers (obligate according to their posi-
tion within a known Lynch syndrome family) on the
registry were identified. Individuals were excluded if
they lived outside the region or were known to have
their surveillance performed outside the region. Individ-
uals who were alive at the time of the study were identi-
fied. To ensure that data were as complete as possible
screening data were gathered or confirmed retrospec-
tively. The hospital at which each patient should be
undergoing colonic surveillance was confirmed from the
medical notes, by contact with the patient by their
genetic counsellor or by contacting the local hospitals.
All colonoscopy reports from the date of Lynch syn-
drome diagnosis were gathered. The date of diagnosis
was counted as the date of a positive Lynch syndrome
mutation test or as the date of clinical diagnosis for
obligate carriers. Histopathology reports of biopsies,
polypectomies and resection specimens were also
sought. When gaps in surveillance or a lack of screening
were identified, further investigation into medical case
notes or direct patient enquiry via the genetic council-
lors was conducted in order to clarify the cause(s) for
noncompliance. The surveillance interval was compared
with the recommended British Society of Gastroenterol-
ogy guidelines [4]. Surveillance colonoscopies were
described as compliant with guidelines if they were con-
ducted within 3 months of the upper limit of 2 years
(27 months). Records were included in analysis of time-
39
HNPCC colonoscopy screening compliance K. Newton et al.

liness of surveillance if the interval from the previous
endoscopy was known. In order to investigate the out-
comes of screening, all endoscopies with complete
reports were included. To assess the incidence of CRC
during surveillance, only CRCs that were diagnosed
after the date of a positive gene mutation test were
included.
The incidence of CRC in the screened population
was compared with the incidence in mutation-positive
patients not undergoing surveillance. In order to esti-
mate CRC incidence in mutation-positive individuals
not undergoing screening, all mutation-positive individ-
uals on the database were identified (n = 689) and can-
cer incidence was censored at the date of family
ascertainment. The date of family ascertainment was
used as a surrogate marker for the date of commence-
ment of surveillance. Therefore, the CRC incidence in
all mutation carriers prior to the date of their family
ascertainment is a surrogate for CRC incidence in
individuals not undergoing screening. This assumes that
individuals were not screened prior to family ascertain-
ment. The time until onset of CRC was calculated
actuarially from date of birth until date of diagnosis of
cancer. If the patient had not been diagnosed with can-
cer at the time of the study, the date of last follow-up
was used and the cancer-free time was predicted using
Kaplan–Meier curves. Cumulative incidence of CRC to
the age of 70 was calculated. Kaplan–Meier curves were
used to estimate mean age of onset of CRC in each
group.
Statistical analysis
Analyses were performed using SPSS 13.0 software
(IBM, Armonk, New York, USA). Kaplan–Meier analy-
sis and log-rank (Mantel–Cox) tests were conducted.
Results
Data on large bowel surveillance were collected for 227
Lynch syndrome mutation carriers or obligates (85
MLH1, 119 MSH2, 21 MSH6, 2 PMS2) who were
screened at 26 hospitals across the northwest of Eng-
land. A complete record of surveillance was available for
183 patients. The most recent one or two colonoscopy
reports were available for the remaining 44 patients. Six
hundred and four endoscopy reports were available.
The median surveillance time was 4.4 (0.1–24) years
(Table 1).
Timeliness
There were 439 endoscopy records available for assess-
ing the timeliness of surveillance. Records were only
included in the analysis of timeliness if the interval from
the previous endoscopy was known.

Table 1 Colorectal cancer (CRC) in a Lynch syndrome screening population during the median surveillance time of 4.4 (0.1–24)
years.

Interval from previous Quality of prep. at Previous

Age (years) Site Dukes stage screen (months) previous screen Gene CRC?

43 Rectum A 8 Unknown MLH1 No

83 Mid transverse A 8 Satisfactory MSH6 Yes
58 Rectum A 9 Very poor MLH1 Yes
66 Anastomosis site A 11 Satisfactory MLH1 Yes
69 Rectum A 15 Good MLH1 Yes
49 Sigmoid A 15 Unknown MLH1 No
53 Rectum A 21 Unknown MSH2 No
42 Rectum A 22 Satisfactory MSH2 Yes
40 Caecum B 24 Unknown MSH2 No
48 Rectum B 24 Unknown MSH2 Yes
47 Ascending C 35 Unknown MSH2 Yes
64 Sigmoid B 36 Unknown MLH1 Yes
45 Caecum B 36 Unknown MLH1 Yes
40 Transverse C 51 Good MLH1 No
34 Ascending B First screen N/A MSH2 No
46 Transverse B First screen N/A MSH2 No
64 Ascending C First screen N/A MSH2 Yes
70 Low rectum/anus D First screen N/A MSH2 Yes
55 Caecum – unconfirmed C First screen N/A MSH2 No

N/A, not applicable.

40 Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
K. Newton et al.
Out of the total of 439 endoscopies, 299 (68.1%)
took place within 27 months of the previous screen or
within 3 months of a positive gene test and 140
(31.9%) were delayed. Of these 140 delayed examina-
tions, 38 (27.1%) were first screenings conducted more
than 3 months after the date that a patient had a posi-
tive mutation test and 21 (15.0%) were delayed due to
incorrect planning, for example at 3 years. Of the 439
screenings, 32 (7.3%) were delayed by more than
12 months. Compliance at a single point in time (1
November 2011) was investigated. Within the previous
27 months, or within 3 months of diagnosis with
Lynch syndrome if it was the first screening, 87.3% of
patients had had a colonoscopy.
Of the 227 Lynch syndrome patients under screen-
ing, 10 were no longer undergoing surveillance. Four
of these had been incorrectly discharged by clinicians at
endoscopy.
Three individuals deviated from the surveillance pro-
tocol by more than 12 months due to pregnancy, non-
response to contact or moving with no screening
arrangement put in place at their new location.
Delays due to hospital system errors include delays
in booking by the endoscopy department and automatic
discharge of the patient if they failed to respond or
attend. Clinician errors include incorrect interval plan-
ning and incorrectly discharging the patient. Patient
issues include cancelling or not attending appointments,
nonresponse to contact from endoscopy departments,
changing address and not informing the hospital of
their new address or delay in setting up screening in
their new location, and being pregnant or breastfeeding.
There was variation in the degree of delay between sur-
veillance centres. A few hospitals had particularly high
proportions of delays. In addition, some patients had
more than one delayed colonoscopy. This may be due
to particular patients who are noncompliant.
Quality of screening
Completeness of examination was known for 339 colo-
noscopies. Of these, 313 (92.3%) were complete and 26
(7.7%) did not reach the caecum. In the latter, the
method of follow-up was by barium enema (eight
cases), repeat colonoscopy (six cases) and CT colonog-
raphy (five cases). The quality of the bowel preparation
was known for 376 colonoscopies. In 340 (90.4%)
bowel preparation was considered good or satisfactory
and poor in 36 (9.6%). In half of these cases, follow-up
was not altered because of poor bowel preparation. In
13 cases, the interval before the next screening was
reduced. In three cases the plan was to repeat the
endoscopy.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
HNPCC colonoscopy screening compliance
Colonoscopy findings
The colonoscopy findings were known for 528 colonos-
copies. Of these, 404 (76.5%) were reported as normal.
A CRC was found in 19 (3.6%) examinations and of
these 11 (57.9%) had had a previous CRC. Eight
occurred in MLH1 mutation carriers, 10 in MSH2
mutation carriers and one in a MSH6 mutation carrier.
At least one adenoma was found during 48 (9.1%) of
the colonoscopies, one of which required bowel resec-
tion. In 32 (6.1%) only hyperplastic polyps were found
and in 24 (4.5%) polyps with unknown histology were
identified. Three CRCs were found during three colo-
noscopies delayed by more than 12 months. A large
adenoma requiring surgery was found in one case where
screening was delayed by more than 12 months
(Fig. 1).
Incidence of colorectal cancer
In the 227 individuals undergoing surveillance, the inci-
dence of CRC to age 70 was 24.9% (Table 2) and in
the 689 not under surveillance it was 81.1% (log rank
Mantel–Cox, P < 0.0001) (Fig. 2). The age of onset of
CRC was significantly greater in the population under-
going surveillance (67.5 years) than in those not under-
going surveillance (52.8 years) (log rank Mantel–Cox,
P < 0.001).
The mean ages at last follow-up in the groups
undergoing and not undergoing surveillance were
51.3 years (median 51 [25–88] years) and 52.3 years
(median 52 [17–93] years), respectively. The propor-
tion of the different genes affected was similar between
the groups. In the group undergoing surveillance there
was no significant difference in CRC incidence
between affected genes (Fig. 3), but in the unscreened
group the CRC incidence was higher in MLH1 and
MSH2 compared with MSH6 carriers (P < 0.01, see
Fig. 4).
Discussion
This is the first study in the UK to investigate compli-
ance with large bowel screening in Lynch syndrome
mutation carriers. We have demonstrated that objective
compliance with screening is 62%. Compliance at a
recent single point in time, which may be a better mea-
sure of current practice, was 87%. It also found the
cumulative incidence of CRC to the age of 70 years to
be reduced by current NHS screening from 81% in
mutation-positive Lynch syndrome patients not under-
going screening to 25% in mutation-positive individuals
undergoing screening (Table 3).
41
HNPCC colonoscopy screening compliance
All pathogenic Lynch syndrome mutation
carriers or obligate carriers
(dead and alive) n = 689
All pathogenic Lynch syndrome mutation
carriers or obligate carriers
(alive at time of study)
CRC incidence censored at date of family
ascertainment (surrogate marker for date
of commencement of screening)
Unscreened population
n = 689
Figure 1 Flow diagram of the study population (CRC, colorectal cancer).
Compliance with screening has been investigated in
other European and American high-risk populations.
In Finland and the Netherlands, patients with Lynch
syndrome are managed through centralized national
registries, which coordinate the genetic investigation,
diagnosis and screening of high-risk individuals. Bleiker
et al. [21] investigated objective and self-reported
screening compliance in the Dutch registry population.
Families with Lynch syndrome (27 families fulfilling
the Amsterdam Criteria [1] ) and high to moderate
risk families (43 families) were included (n = 178).
Among these, 184 individuals responded to the ques-
tionnaire, of whom 91 (61%) had Lynch syndrome
and 31 (16.8%) were mutation carriers. Overall, 72%
of patients were found objectively to be fully compliant
with screening recommendations. Screening was
delayed by over a year in 20 individuals. Thirteen had
only had one screen, and four had never been
screened. In the Lynch syndrome mutation carrier
group, 86% were fully compliant. Median follow-up
was 6 years. The reasons for noncompliance were per-
ceived discomfort and embarrassment at the time of
the procedure and human error. Pylvainainen et al.
[22] reported 98% compliance with 3-yearly screening
of 587 Lynch syndrome mutation carriers coordinated
through the Finnish registry. Causes for delayed or
interrupted screening were pain and discomfort associ-
ated with the procedure. The much higher compliance
with screening in both the Finnish and Dutch popula-
tions could be attributed to effective communication
and counselling facilitated by the registry. Studies from
the United States have demonstrated much poorer
42 Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
K. Newton et al.
n = 267
Excluded-pathogenic
Lynch syndrome
mutation carriers or
obligate carriers (alive
at time of study) who
live or are screened
out of region
n = 40
Screened population
n = 227
compliance. Stoffel et al. [20] conducted a survey-
based study of families with Lynch syndrome (fulfilling
the Amsterdam criteria). Of 468 eligible patients, only
270 (58%) completed the questionnaire and 73% of
patients were found to have had screening within the
guidelines of 2 years. As the number of responders was
only 58%, however, it is likely that real compliance was
lower.
The current standard of care for bowel cancer
screening in the UK is the National Bowel Cancer
Screening Programme (BSCP). This offers biannual
screening (faecal occult blood testing followed by colo-
noscopy) to individuals in the population at risk from
the ages of 60 to 74 years. Strict quality assurance mea-
sures are rigorously reported and include caecal intuba-
tion rate (95%), withdrawal time (6 min), adenoma
detection rate (30%), polyp retrieval rate (90%) and
quality of bowel preparation (excellent or adequate in
95%) [23]. In our study the caecal intubation rate was
92.3%, below the 95% standard required in the BCSP.
The adequate bowel preparation rate was 90.4% which
again was below the 95% required by the BCSP. Ade-
noma detection rates were 9% in this study. This is not
comparable to the BSCP data as the risk is very differ-
ent. Dove-Edwin et al. [24] reported a rate of adenoma
detection of 15.4% at first colonoscopy in a group of
families fulfilling the Amsterdam Criteria and mutation-
positive individuals. In the only controlled trial of
screening, Jarvinen et al. [2] reported adenomas in 15/
44 mutation carriers over a 15-year period of 3-yearly
screening. This corresponds to an adenoma detection
rate of 6.8%.
K. Newton et al. HNPCC colonoscopy screening compliance

(57.7–64.3)
Kaplan–meier curve demonstrating cumulative

No cancers

No cancers
screening

screening
incidence of CRC in Lynch Syndrome mutation

during

during
MSH6
carriers in patients undergoing screening compared

61.0
to those not undergoing screening
1.0
Log rank mantel cox P < 0.001

(30.4–60.2)
No cancers

No cancers
screening

screening
Chi square 116.7

during

during

Cumulative incidence of CRC

0.8 Df 1
PMS2

45.3
Population not
0.6 undergoing screening
(66.2–69.3)

(67.0–69.8)

(51.5–55.0)
Screening population
Population not undergoing
MSH2

screening-censored
67.3

68.4

53.2
0.4 Screening population-
in years (95% confidence interval)

censored
(64.3–68.8)

(64.7–69.1)

(48.7–52.4)
Mean age of onset of CRC

0.2
MLH1

66.5

66.9

50.5

0.0
0 10 20 30 40 50 60 70
(66.4–68.7)

(66.9–69.1)

(51.6–54.0)

Time to CRC (years)

Figure 2 Kaplan–Meier curve demonstrating the cumulative
67.5

68.9

52.8

incidence of colorectal cancer (CRC) in Lynch syndrome muta-

Table 2 Kaplan–Meier risk estimate (1-KM) of cumulative incidence of colorectal cancer (CRC) to age 70 years.

All

tion carriers undergoing screening compared with those not

undergoing screening.
(51.5–70.3%)
No cancers

No cancers
screening

screening
during

during
MSH6

Kaplan–meier curve demonstrating cumulative

60.9%

incidence of CRC in Lynch Syndrome mutation

carriers undergoing screening
(33.4–77.8%)

1.0 Mutation
Log rank mantel cox
No cancers

No cancers

MLH1
screening

screening

All MSH2
during

during

Cumulative incidence of CRC

P = 0.311
55.6%

PMS2
PMS2

0.8 MSH6
Chi square 3.58
MLH1-censored
df 3 MSH2-censored
PMS2-censored
(12.0–33.4%)

(78.9–87.7%)

0.6 MLH1 vs MSH2 MSH6-censored

(4.0–10.2%

P = 0.386
Chi square 0.751
MSH2

23.2%

7.1%

83.3%

0.4 Df 1
(21.7–52.5%)

(12.0–27.8%)

(81.4–88.6%)

0.2
1-KM risk of CRC to age 70
(95% confidence interval)

MLH1

37.1%

19.9%

85.0%

0.0
20 30 40 50 60 70
Time to CRC (years)
(17.1–32.4%)

(10.6–23.8%)

(78.4–83.8%)

Figure 3 Kaplan–Meier curve demonstrating the cumulative

incidence of colorectal cancer (CRC) in Lynch syndrome muta-
24.9%

17.2%

81.1%

tion carriers undergoing screening.

All

Population not being

Screening population

Screening population

screened (n = 689)

There are some limitations to this study. This was

cancers diagnosed

not randomized and there may therefore be inherent

on first screen

differences between the screened and unscreened popu-

(n = 227)

excluding

lations which may have affected the incidence of CRC.

We used the incidence of CRC in all mutation carriers
on the registry as a surrogate for CRC in unscreened

Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49 43
HNPCC colonoscopy screening compliance K. Newton et al.

Kaplan–meier curve demonstrating cumulative incidence of CRC

in Lynch Syndrome mutation carriers not undergoing screening

1.0 Log rank mantel cox Gene Mutation

MLH1
MSH2
MLH1 vs MSH2 PMS2
MSH6

Cumulative incidence of CRC

0.8 P = 0.055, Chi square 3.69, Df 1
MLH1-censored
MSH2-censored
MLH1 vs MSH6 PMS2-censored
0.6 P < 0.001, Chi square 17.24, Df 1 MSH6-censored

MLH2 vs MSH6
0.4 P < 0.001, Chi square 11.9, Df 1

0.2

0.0
0 10 20 30 40 50 60 70
Time to CRC (years)
Figure 4 Kaplan–Meier curve demonstrating the cumulative incidence of colorectal cancer (CRC) in Lynch syndrome mutation
carriers not undergoing screening.

Table 3 Kaplan–Meier risk estimate (1-KM) of cumulative incidence of colorectal cancer (CRC) by age.

1-KM Cumulative Incidence of CRC by age(95% confidence interval)

< 30 years < 40 years < 50 years < 60 years < 70 years

Screening All mutation carriers 0.4% 1.0% 6.3% 8.6% 24.9%

population (MLH1, MSH2, PMS2, (0–0.8%) (0.3–1.7%) (4.4–8.2%) (6.1–11.1%) (17.1–32.7%)
MSH6) (n = 227)
MLH1 mutation 1.2% 2.6% 7.7% 11.5% 37.1%
carriers (n = 85) (0–2.4%) (0.8–4.4%) (5.3–10.1%) (6.6–16.4%) (21.7–52.5%)
MSH2 mutation No cancers 1.1% 6.5% 9.8% 23.4%
carriers (n = 119) (0–2.2.0%) (3.7–9.3%) (7.6–12.0%) (12.2–22.4%)
Population not All mutation carriers 4.4% 21.8% 41.5% 65.9% 81.1%
being screened (MLH1, MSH2, PMS2, (3.6–5.2%) (20.0–23.6%) (39.3–43.7%) (63.3–68.5%) (83.8–88.3%)
MSH6) (n = 689)
MLH1 mutation 6.8% 26.4% 49.4% 73.1% 85.0%
carriers (n = 285) (5.2–12.0%) (23.5–29.3%) (45.9–52.9%) (69.5–76.7%) (81.4–88.6%)
MSH2 mutation 2.8% 20.2% 38.5% 67.0% 82.3%
carriers (n = 338) (1.8–3.8%) (17.7–22.7%) (35.3–41.7%) (63.1–70.9%) (77.9–86.7%)

individuals to minimize this effect. The screened popu-
lation includes all mutation-positive individuals living in
the geographical area of the registry who where alive at
the start of the study. This may have favourably affected
the incidence of CRC in the screened population
because those with cancers occurring at a younger age
may have died after family identification before the
beginning of the study. However, the groups were of
similar age and there was a similar distribution of the
genes affected. There was also a possibility that the
cumulative incidence of CRC reflected ascertainment
bias since the patients were identified through clinics
selecting individuals at high risk.
The estimated cumulative incidence of CRC of
81.1% to age 70 in Lynch syndrome carriers not being
screened is consistent with previous studies. Plaschke
et al. [25] examined mutation-positive individuals from
706 families on the German HNPCC Registry and
found a cumulative lifetime incidence of 90% for MLH1
and MSH2 carriers. Aarnio et al. [26] estimated the

44 Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
K. Newton et al.
cumulative incidence to age 70 years of CRC in 360
mutation-positive individuals on the Finnish HNPCC
Registry to be 82%. There was no significant difference
in the cumulative incidence of CRC between individuals
with MLH1 or MSH2 mutations, consistent with previ-
ous studies [27–29], whether undergoing screening or
not. MSH6 mutation carriers have a significantly lower
cumulative incidence of CRC than those with either
MLH1 or MSH2. This is consistent with previous stud-
ies [25]. Again this is true both with and without
screening.
The present study has demonstrated that surveillance
of high-risk individuals according to the BSG guidelines
greatly reduces the incidence of CRC to age 70. By
excluding the cancers diagnosed on first screen it is pos-
sible to get a more accurate picture of the incidence of
CRC in screened mutation-positive individuals. When
this was done, the incidence of CRC was 17.2% to age
70. It is noteworthy that the median surveillance time
was only 4.4 years, an average of about two colonos-
copies per individual, indicating that the effect of sur-
veillance is strongly evident even in such a short time
frame.
The only previous study of surveillance in a high-risk
population in the UK was conducted by Dove-Edwin
et al. [24] in 2005. This was a prospective observational
study of high-risk families over a 16-year period and
individuals were offered colonoscopy every 5 years (or
3-yearly if adenomas were detected). Families were cate-
gorized simply by family history and cancer incidence,
and survival was compared between those who accepted
screening and those who did not. The authors found
that in individuals meeting the Amsterdam Criteria the
incidence of CRC was half that in the screened than in
the non-screened group.
The benefits of large bowel screening in patients
with Lynch syndrome are clear. Whilst screening is asso-
ciated with a large reduction in the incidence of CRC,
this is still much greater than in the general population.
Delays in screening are caused by problems with book-
ing in endoscopy departments, patients being discharg-
ing inappropriately from follow-up, incorrect planning
of screening intervals, patients cancelling appointments
or not attending or failing to communicate a change of
address. These problems could be reduced by an effec-
tive coordinated screening programme.
Of the 19 CRCs diagnosed in the surveillance group,
five were on first screen. This also highlights the need
for urgent surveillance of family members following
family ascertainment. Of these 19, 11 were found in
patients who had previously had a CRC. Surveillance
should therefore be aimed particularly at individuals
who have had a previous cancer. Cancer incidence and
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
HNPCC colonoscopy screening compliance
survival in Lynch syndrome may be further improved by
implementing regional coordinated surveillance pro-
grammes with the same quality assurance and audit pro-
cesses as the NBCSP. Each region should have a
designated surveillance coordinator to facilitate commu-
nication between screening centres, patients and the
genetic medicine department.
The incidence of CRC in Lynch syndrome is reduced
by surveillance, and the age of onset of CRC is 15 years
later. This study provides compelling evidence that
administration and surveillance of patients with Lynch
syndrome should be performed to the same high stan-
dards as required for the UK NBCSP.
Author contributions
Study conception and design: KN, KG, JH. Acquisition
of data: KN, KG. Analysis and interpretation of data:
KN, KG, FL, DGE, JH. Writing manuscript: KN, KG,
FL, DGE, JH.
References
1 Lynch HT, de la Chapelle A. Hereditary colorectal cancer.
N Engl J Med 2003; 348: 919–32.
2 Jarvinen HJ, Aarnio M, Mustonen H et al. Controlled
15-year trial on screening for colorectal cancer in families
with hereditary nonpolyposis colorectal cancer. Gastroenter-
ology 2000; 118: 829–34.
3 Jass JR. Classification of colorectal cancer based on correla-
tion of clinical, morphological and molecular features. His-
topathology 2007; 50: 113–30.
4 Cairns SR, Scholefield JH, Steele RJ et al. Guidelines for
colorectal cancer screening and surveillance in moderate
and high risk groups (update from 2002). Gut 2010; 59:
666–89.
5 Vasen HF, Mecklin JP, Watson P et al. Surveillance in hered-
itary nonpolyposis colorectal cancer: an international cooper-
ative study of 165 families. The International Collaborative
Group on HNPCC. Dis Colon Rectum 1993; 36: 1–4.
6 Mecklin JP, Jarvinen HJ, Aukee S, Elomaa I, Karjalainen K.
Screening for colorectal carcinoma in cancer family syn-
drome kindreds. Scand J Gastroenterol 1987; 22: 449–53.
7 Love RR, Morrissey JF. Colonoscopy in asymptomatic indi-
viduals with a family history of colorectal cancer. Arch
Intern Med 1984; 144: 2209–11.
8 de Jong AE, Hendriks YM, Kleibeuker JH et al. Decrease
in mortality in Lynch syndrome families because of surveil-
lance. Gastroenterology 2006; 130: 665–71.
9 Vasen HF, den Hartog Jager FC, Menko FH, Nagengast
FM. Screening for hereditary non-polyposis colorectal can-
cer: a study of 22 kindreds in The Netherlands. Am J Med
1989; 86: 278–81.
10 de Vos tot Nederveen Cappel WH, Nagengast FM,
Griffioen G et al. Surveillance for hereditary nonpolyposis
45
HNPCC colonoscopy screening compliance
colorectal cancer: a long-term study on 114 families. Dis
Colon Rectum 2002; 45: 1588–94.
11 Vasen HF, Nagengast FM, Khan PM. Interval cancers in
hereditary non-polyposis colorectal cancer (Lynch syn-
drome). Lancet 1995; 345: 1183–4.
12 Renkonen-Sinisalo L, Aarnio M, Mecklin JP, Jarvinen HJ.
Surveillance improves survival of colorectal cancer in
patients with hereditary nonpolyposis colorectal cancer.
Cancer Detect Prev 2000; 24: 137–42.
13 Dunlop MG. Guidance on gastrointestinal surveillance for
hereditary non-polyposis colorectal cancer, familial adeno-
matous polypolis, juvenile polyposis, and Peutz-Jeghers
syndrome. Gut 2002; 51(Suppl 5): V21–7.
14 Jarvinen HJ, Mecklin JP, Sistonen P. Screening reduces
colorectal cancer rate in families with hereditary nonpolyp-
osis colorectal cancer. Gastroenterology 1995; 108:
1405–11.
15 Hadley DW, Jenkins JF, Dimond E, de Carvalho M,
Kirsch I, Palmer CG. Colon cancer screening practices
after genetic counseling and testing for hereditary
nonpolyposis colorectal cancer. J Clin Oncol 2004;
22: 39–44.
16 de Leon MP, Pedroni M, Benatti P et al. Hereditary
colorectal cancer in the general population: from cancer
registration to molecular diagnosis. Gut 1999; 45: 32–8.
17 Jass JR. Screening for hereditary bowel cancer in New Zealand.
Gut 1994; 35: 1328.
18 Mecklin JP, Jarvinen HJ. Surveillance in Lynch syndrome.
Fam Cancer 2005; 4: 267–71.
19 Ponz de Leon M, Benatti P, Di Gregorio C et al. Genetic
testing among high-risk individuals in families with heredi-
tary nonpolyposis colorectal cancer. Br J Cancer 2004; 90:
882–7.
20 Stoffel EM, Mercado RC, Kohlmann W et al. Prevalence
and predictors of appropriate colorectal cancer surveillance
Commentary on ‘Colonoscopy screening compliance and
outcomes in patients with Lynch syndrome’ doi:10.1111/codi.12841
Findings of the first family with syndromic gastrointesti-
nal cancer were published in 1966 [1]. Since then,
Lynch syndrome, sometimes known as hereditary non-
polyposis colorectal cancer, has been recognized as the
commonest hereditary cancer syndrome affecting
the development of 2–5% of all colorectal cancers [2].
The implications of the molecular genetic diagnosis of
this syndrome are immense – to the individual, family
and society, and, as a whole, they harbour lifesaving
issues! Colorectal cancer that develops on a background
of mismatch repair (MMR) gene mutation bears unique
phenotypic characteristics with the potential to impact
46 Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 17, 38–49
K. Newton et al.
in Lynch syndrome. Am J Gastroenterol 2010; 105: 1851–
60.
21 Bleiker EM, Menko FH, Taal BG et al. Screening behavior
of individuals at high risk for colorectal cancer. Gastroenter-
ology 2005; 128: 280–7.
22 Pylvanainen K, Kairaluoma M, Mecklin JP. Compliance
and Satisfaction with Long-Term Surveillance in Finnish
HNPCC Families. Fam Cancer 2006; 5: 175–8.
23 http://info.cancerresearchuk.org/cancerstats/types/bowel/
Bowel. Cancer Research UK. (accessed 1 November 2011).
24 Dove-Edwin I, Sasieni P, Adams J, Thomas HJ. Prevention
of colorectal cancer by colonoscopic surveillance in individ-
uals with a family history of colorectal cancer: 16 year, pro-
spective, follow-up study. BMJ 2005; 331: 1047.
25 Plaschke J, Engel C, Kruger S et al. Lower incidence of
colorectal cancer and later age of disease onset in 27 fami-
lies with pathogenic MSH6 germline mutations compared
with families with MLH1 or MSH2 mutations: the German
Hereditary Nonpolyposis Colorectal Cancer Consortium.
J Clin Oncol 2004; 22: 4486–94.
26 Aarnio M, Sankila R, Pukkala E et al. Cancer risk in muta-
tion carriers of DNA-mismatch-repair genes. Int J Cancer
1999; 81: 214–8.
27 Barrow E, Alduaij W, Robinson L et al. Colorectal cancer
in HNPCC: cumulative lifetime incidence, survival and
tumour distribution. A report of 121 families with proven
mutations. Clin Genet 2008; 74: 233–42.
28 Aarnio M, Mecklin JP, Aaltonen LA, Nystrom-Lahti M,
Jarvinen HJ. Life-time risk of different cancers in hereditary
non-polyposis colorectal cancer (HNPCC) syndrome. Int J
Cancer 1995; 64: 430–3.
29 Lin KM, Shashidharan M, Thorson AG et al. Cumulative
incidence of colorectal and extracolonic cancers in MLH1
and MSH2 mutation carriers of hereditary nonpolyposis
colorectal cancer. J Gastrointest Surg 1998; 2: 67–71.
therapy. A well-informed healthcare team must acquaint
itself with the patient’s family history in order to recog-
nize any familial and/or colonoscopic pattern of colo-
rectal cancer occurrence. Reflex testing of a patient who
might fit clinical criteria under revised Bethesda guide-
lines [3] has been practised to date but may not be
completely effective in the detection of Lynch syndrome
[4]. Universal screening of all colorectal cancer patients
for Lynch syndrome is therefore recommended [5] and
shown to be cost effective [6]. Programmes across the
USA are evenly split in screening practices [7]. In either
case, the real benefit of diagnosing MMR gene muta-