PURPOSE: Similar to findings obtained for most carcinomas, the pathogenesis
of colorectal cancer is considered to be multifactorial. There is strong e
vidence for an inherited, genetic predisposition to disease in patients wit
h familial adenomatous polyposis and hereditary nonpolyposis colorectal can
cer. There is still debate, however, about the contribution of genetic fact
ors to the pathogenesis of sporadic colorectal cancer. The present study wa
s undertaken to search for human leukocyte antigen associations in a group
of patients with colorectal cancer and to correlate the findings with both
the histology of the disease and family history. SUBJECTS AND METHODS: The
allele frequencies of serologically defined human leukocyte antigen class I
and II antigens were studied in 101 patients with a recent, histologically
confirmed diagnosis of colorectal cancer. AU individuals in this study wer
e unrelated to each other. After surgical treatment, all patients were grou
ped according to the stage (Dukes Stages A, B, C, and D), differentiation (
Grades 1, 2, and 3), and the site of the tumor. Patients were also classifi
ed with regard to family history for colorectal cancer. The results obtaine
d for human leukocyte antigen frequencies were compared with those of 105 h
ealthy control subjects (control group). RESULTS: An increased frequency of
human leukocyte antigen-B18 (27.72 vs. 14.28 percent; P < 0.025; odds rati
o = 2.3) and of human leukocyte antigen-DQ5 (43.56 vs. 22.5 percent; P < 0.
01; odds ratio = 2.65) was observed for patients with colorectal cancer vs.
control subjects, respectively. In addition, human leukocyte antigen-B18 w
as present with increased frequency (30.76 percent; P < 0.05; odds ratio =
2.66; and 26.67 percent; P < 0.05; odds ratio = 2.18) among patients with r
ectal and colon carcinoma, respectively. A higher frequency of human leukoc
yte antigen-DQ5 (45.33 percent; P < 0.01; odds ratio = 2.84) was observed a
mong patients with colon carcinoma. Remarkably, human leukocyte antigen-DQ5
(50 vs. 22.5 percent; P < 0.05; odds ratio = 3.43) and human leukocyte ant
igen-Al (41.66 vs. 12.38 percent; P < 0.01; odds ratio = 5.05) were found t
o be strongly associated with a family history of colorectal cancer. CONCLU
SION: The observation of specific human leukocyte antigen associations with
particular subsets of colorectal cancer strongly suggests that genetic sus
ceptibility for the development of colorectal cancer exists. Although the m
ultifactorial pathogenesis of colorectal cancer must be considered, human l
eukocyte antigens may have useful predictive and diagnostic value.