A Prospective Study of the Clinical, Genetic,
Screening, and Pathologic Features of a Family
With Hereditary Mixed Polyposis Syndrome
P. Rozen, M.B., B.S., Z. Samuel, and E. Brazowski, M.D.
Departments of Gastroenterology and Pathology, Tel Aviv Medical Center; and Tel Aviv University, Tel Aviv,
In 1997, hereditary mixed polyposis syndrome (HMPS) was
described in an Ashkenazi pedigree having colorectal pol-
yps with mixed histology and risk for colorectal cancer
(CRC). The mutation is now localized to 15q13–14. Since
1980, compliant relatives of an HMPS family were seen
annually, tested genetically, and had colonoscopy offered
every 1 to 2 yr from age 20 yr. The Israeli pedigree has 37
members (17 clinically affected by CRC or polyps), and
seven of 13 available relatives entered our screening pro-
gram. The others, followed-up elsewhere, provided clinical
information. Half of our screened group had rectal bleeding;
others were asymptomatic. Colonoscopy, performed a mean
of four times, identified polyps in all seven patients (mean
age 28 yr). Polyps were removed and included juvenile
adenomas, mixed juvenile adenomas, hyperplastic polyps,
mixed hyperplastic adenomas, serrated adenomas, and tu-
bular adenomas. None of our screened patients developed
CRC or extracolonic neoplasia. Linkage analysis localized
their mutation to 15q13–14. This high-penetrance founder
mutation so far is described only in Ashkenazim. The CRC
pathway seems to be through juvenile and hyperplastic
polyps. Mutation identification will aid screening for and
evaluation of HMPS prevalence in Jewish and non-Jewish
populations. Meanwhile, a cancer pedigree and correct clas-
sification of polyps will identify HMPS families. They re-
quire early and frequent colonoscopy, polypectomy, and
elective extensive colectomy when indicated. (Am J Gas-
troenterol 2003;98:2317–2320. © 2003 by Am. Coll. of
In 1989, Murday and Slack described a family with mixed
large bowel polyps and increased risk for colorectal cancer
(CRC) (1). In 1997, Whitelaw et al. (2) defined the disorder
as an autosomal dominant syndrome characterized by an
increased risk for CRC, but not for extracolonic cancer, and
having large bowel polyps (?15 in number at initial exam-
ination) of differing histology. These included atypical ju-
venile polyps, hyperplastic polyps, and adenomas having
mixed histologic features of both types. Whitelaw et al.
named this hereditary mixed polyposis syndrome (HMPS).
Since then, it has been recognized that this family is a large
international Jewish pedigree of Lithuanian origin, and the
mutation has now been localized by linkage to the area of
We identified the Israeli HMPS relatives; they have been
cooperative, providing demographic and clinical data on
their extended family, and since 1980 they have participated
in a screening program. This provided an opportunity for us
to study for an extended period and summarize the spectrum
of clinical and histologic features of HMPS.
From the age of 20 yr, at-risk relatives were invited for
annual clinical evaluation and were offered colonoscopy
and polypectomy every 1 to 2 yr. All available histologic
material was reviewed and classified according to recog-
nized criteria (4). At-risk relatives were invited to partici-
pate in a multinational study to localize and identify the
responsible mutation (3).
In Israel, 37 HMPS relatives were identified, 17 of whom
were clinically affected by CRC or polyps (Fig. 1). CRC
was reported to have occurred in generations 1–3 at the ages
of 40–65 yr. Benign polyps have been identified in the more
recent generations (4 and 5) under our clinical observation,
and no fatal CRC has occurred. Two relatives followed-up
elsewhere had an elective subtotal colectomy for polyposis
at age 47 yr; one was reoperated successfully at age 57 yr for
CRC that developed in the remaining large bowel.
Colonoscopy examinations have been performed a mean
3.8 ? 2.4 (SD) times in the screened relatives. The mean
age at first detection of a colorectal polyp was 28 yr, with
a range of 10–47 yr. Half of the relatives were asymptom-
atic, and the others had some rectal bleeding. Elective gas-
troscopy was performed on only one asymptomatic patient
with no abnormal findings, and we have not continued to
advise its performance. No extracolonic neoplasia has
THE AMERICAN JOURNAL OF GASTROENTEROLOGY
© 2003 by Am. Coll. of Gastroenterology
Published by Elsevier Inc.
Vol. 98, No. 10, 2003
The number of polyps detected at any examination ranged
from one to 23. One was 1.5 cm in diameter, two were
0.5–0.9 cm, and the others were ?0.5 cm. The cumulative
number of polyps removed was 54; one in two patients, two
in two patients, and four, nine, and 38, respectively, in the
Review of the available histologic material allowed us to
classify the polyps as follows: juvenile polyps, 10; mixed
Figure 2. Mixed juvenile–adenomatous polyp. (A) Overview of the polyp (hematoxylin and eosin stain, ?40, original magnification).
Different areas of the same polyp composed of (B) serrated adenomatous crypts (?100, original magnification) and (C) dilated
nondysplastic crypts consistent with juvenile polyp (?100, original magnification).
Figure 1. Extended pedigree of the HMPS family in Israel. Almost half the family has had colorectal cancer or polyps.
2318Rozen et al.
AJG – Vol. 98, No. 10, 2003
juvenile–adenomatous polyps, 21 (Fig. 2); hyperplastic pol-
yps, 8; mixed hyperplastic–adenomatous polyps, 1 (Fig. 3);
serrated adenomas, 12 (Fig. 4); and tubular adenomas, 2.
The polyp numbers, size, and adenomatous elements in-
creased with age of the patients.
All family members contributed DNA for genetic re-
search studies performed at the Molecular and Population
Genetics Laboratory, Cancer Research, London. Their mu-
tation site was localized in 15q13–14 (3).
In these at-risk relatives, HMPS was identified only by a
careful creation of the cancer pedigree and by recognizing
the significance of the histologic classification of their pol-
yps. To date, careful screening and preventive polypectomy
have prevented CRC and/or mortality from CRC. In our
screened population, the number of polyps found at initial
colonoscopy was usually sparse and rarely demonstrated
high-grade dysplasia. This could be owing to the young age
of initiating surveillance and an aggressive policy of early
and repeated polypectomy.
The HMPS mutation has been localized to 15q13–14 (3).
Other unrelated Ashkenazi families in the United Kingdom
and Israel have also been identified as having the same
linkage (3, 5). These families have benign adenomas of
various histology and CRC but also noncolonic cancers,
including pancreatic, renal, and breast tumors. A similar
clinical syndrome has been described in a non-Jewish fam-
ily, but they have not been evaluated genetically (6).
The pathway of carcinogenesis seems to be through the
juvenile/hyperplastic polyp, mixed/serrated adenoma–can-
cer sequence (7). However, this can only be confirmed by
molecular studies that have not yet been performed (8, 9).
Ashkenazi Jews are at increased risk for CRC, probably
because of a number of founder genetic perturbations. These
include the low penetrance I130K APC variant (10) and
heterozygote carriers of Bloom syndrome (11). These in-
crease the risk for CRC by only 1.5–2-fold. The high pen-
etrance loci include the MSH2*1906G 3 C mutation (12)
and also the HMPS mutation, which is likely to be a founder
In conclusion, until the exact HMPS mutation is identi-
fied, it is not possible to evaluate its contribution to CRC
risk in the Ashkenazi, non-Ashkenazi, and non-Jewish pop-
ulations. In the meantime, the risk for CRC can be managed
clinically by the careful creation of a cancer pedigree; cor-
rect identification and classification of polyps; and early and
systematic colonoscopy, polypectomy, and elective colec-
tomy when indicated.
Figure 3. Mixed hyperplastic–adenomatous polyp. (A) Overview the polyp (hematoxylin and eosin stain, ?100, original magnification).
(B) Adenomatous part of the polyp with crypts lined by dysplastic epithelium (?200, original magnification). (C) Hyperplastic part shows
crypts with stellate lumens lined by a nondysplastic epithelium (?200, original magnification).
AJG – October, 2003
Family With Hereditary Mixed Polyposis Syndrome
The Family Cancer Clinic is supported by the Israel Cancer
Association. Our thanks to Drs. I. Tomlinson and E. Jaeger,
London, and Dr. J. Jass, Montreal, for their cooperation and
Reprint requests and correspondence: Paul Rozen, M.B., B.S.,
Tel Aviv Medical Center, Department of Gastroenterology, 6
Weizmann Street, Tel Aviv 64239, Israel.
Received Apr. 17, 2003; accepted June 27, 2003.
1. Murday V, Slack J. Inherited disorders associated with colo-
rectal cancer. Cancer Surv 1989;8:139–57.
2. Whitelaw SC, Murday VA, Tomlinson IPM, et al. Clinical and
molecular features of the hereditary mixed polyposis syn-
drome. Gastroenterology 1997;112:327–34.
3. Jaeger EEM, Woodford-Richens KL, Lockett M, et al. An
ancestral Ashkenazi haplotype at the HMPS/CRAC1 locus on
15q13-q14 is associated with hereditary mixed polyposis syn-
drome. Am J Hum Genet 2003;72:1261–7.
4. Fenoglio-Preiser CM, Noffsinger AE, Stemmermann GN, et
al. Gastrointestinal pathology. An atlas and text. Philadelphia:
Lippincott Raven, 1999.
5. Tomlinson I, Rahman N, Frayling I, et al. Inherited suscepti-
bility to colorectal adenomas and carcinomas: Evidence for a
new predisposition gene on 15q14-q22. Gastroenterology
6. Morrison PJ, Chu CE, Nevin NC. A second family with
probable CRAC (colorectal adenomata and carcinoma) syn-
drome—a new familial cancer. Dis Markers 1999;15:120–2.
7. Jass JR, Whitehall VLJ, Young J, et al. Emerging concepts in
colorectal neoplasia. Gastroenterology 2002;123:862–76.
8. Sawyer EJ, Cerar A, Hanby AM, et al. Molecular character-
istics of serrated adenomas of the colorectum. Gut 2002;51:
9. Jass JR. Serrated adenoma of the colorectum. Curr Diagn
10. Rozen P, Naiman T, Strul H, et al. Clinical and screening
implications of the I1307K APC variant in Israeli Ashkenazi
Jews. Evidence for a founder effect. Cancer 2002;94:2561–8.
11. Gruber SB, Ellis NA, Rennert G, et al. BLM heterozygosity
and the risk of colorectal cancer. Science 2002;297:2013.
12. Foulkes WD, Thiffault I, Gruber SB, et al. The founder mu-
tation MSH2*1906G 3 C is an important cause of hereditary
nonpolyposis colorectal cancer in the Ashkenazi Jewish pop-
ulation. Am J Hum Genet 2002;71:1395–412.
Figure 4. Serrated adenoma. Note the dysplastic epithelial lining of the serrated crypts and mucosal surface (hematoxylin and eosin stain,
?200, original magnification).
2320 Rozen et al.
AJG – Vol. 98, No. 10, 2003