AGE AND RISK-FACTORS FOR COLON-CANCER (UNITED-STATES AND AUSTRALIA) -ARE THERE IMPLICATIONS FOR UNDERSTANDING DIFFERENCES IN CASE-CONTROL AND COHORT STUDIES
Ml. Slattery et al., AGE AND RISK-FACTORS FOR COLON-CANCER (UNITED-STATES AND AUSTRALIA) -ARE THERE IMPLICATIONS FOR UNDERSTANDING DIFFERENCES IN CASE-CONTROL AND COHORT STUDIES, CCC. Cancer causes & control, 5(6), 1994, pp. 557-563
Data from two population-based case-control studies were used to inves
tigate the effect of age on colon cancer risk. Dietary intake data wer
e assessed from a study conducted in Utah (United States) between 1979
and 1983; reproductive data were assessed from a study conducted in A
delaide (Australia) between 1979 and 1980. Data from both studies were
assessed for their impact on those less than 65 years of age and thos
e 65 or more years of age. Intake of energy, fat, and protein had a gr
eater impact on risk among older men than among younger men. Risk esti
mates for the upper quartile of intake relative to the lowest quartile
of intake were 8.5 (95 percent confidence interval [CI] = 1.7-43.0) f
or energy, 8.2 (CI = 1.6-41.3) for protein, and 7.2 (CI = 1.6-31.4) fo
r total fat for older men, while comparable risk estimates were 2.4 (C
I = 0.6-9.1) for energy, 3.0 (CI = 0.7-13.6) for protein, and 1.9 (CI
= 0.5-7.1) for total fat among younger men. Similar trends were seen f
or older women for energy and protein. beta-carotene decreased colon c
ancer risk among younger men (odds ratio [OR] = 0.4, CI = 0.1-1.2) and
women (OR = 0.1, CI = 0.1-0.5), although not among older men (OR = 1.
2, CI = 0.3-4.9) and women (OR = 1.9, CI = 0.6-64). Calcium decreased
risk of colon cancer among older men (OR = 0.1, CI = < 0.1-0.8) and yo
unger women (OR = 0.2, CI = < 0.1-0.7). Women who were diagnosed at ag
e 65 or older and were nulliparous had a tenfold increase in colon can
cer risk (CI = 2.4-47.9) relative to women who had an early age at fir
st birth. Women diagnosed with colon cancer before age 65 did not expe
rience an increase of colon cancer risk associated with being nullipar
ous. These data suggest that age at diagnosis may interact with other
factors to alter risk of colon cancer.