in cancer incidence
discussions about time trends in the burden of cancer are full of
debates about apples and oranges. Some people focus on changes in
the incidence of cancer--the proportion of people struck by cancer--
while others focus on cancer mortality, the rate at which deaths
are caused by cancer. These are very different measurements.
are legitimate reasons to focus on cancer mortality, especially
to evaluate the efficacy of cancer treatments. Cancer
mortality rates are vital. If you want to learn, however, about
the causes of cancer and use that to guide cancer prevention, it
is the wrong thing on which to focus. This is because cancer treatment
has improved enormously over the past several decades and thus people
survive much longer now after contracting many types of cancer.
The rates at which people contract cancer can be increasing even
as cancer mortality rates can be going down.
GE, DM Umbach, AJ Sasco, DG Hoel and DL Davis. 1999. Unexplained
increases in cancer incidence in the United States from 1975 to
1994. Annual Review of Public Health 20: 173-209.
et al. examine US data from 1975 to 1994 focusing on long-term
time trends in the incidence of various cancers. They
use a sophisticated statistical analysis to tease apart the effects
of age, period and birth cohort. They
find that while several important cancers are decreasing in frequency,
especially lung cancer, others are increasing.
cancer rates have increased by 22% and 56% among white women and
white men, repectively, over the course of a single generation.
Increases in black men and women are comparable.
contemporary woman's risk of breast cancer is 54% greater than
was her mother's at the same age among blacks and 41% greater
today have about a three- to four-fold risk of being diagnosed
with prostate cancer compared with their fathers."
cancers linked to smoking, or where trends are confounded by changes
in diagnostic procedure (breast and prostate; see below), "relative
to the previous generation, rates increased on average 13% in
black women, 52% in white men, and 67% in black men." There
was little change in white women.
non-Hodgkin's lymphoma, which they analyzed separately, the rates
today relative to 25 years ago have "almost doubled in white
women, nearly tripled in black women, more than tripled in white
men, and more than quadrupled in black men."
H. 1998. Trends in sex-ratio, testicular cancer and male reproductive
hazards: Are they connected? APMIS 106:232-239.
several other authors, Møller reports very significant increases
in testicular cancer among men in northern Europe, more than doubling
in several countries since 1960. His analysis suggests it is linked
to a syndrome of disorders in the male reproductive tract, resulting
from developmental disturbances in the womb. More...
JM, TW Hensle and H Fisch. 2000. Increasing risk of developing
testicular cancer by birth cohort in the United States. Dialogues
in Pediatric Urology 23(1): 7-8.
McKiernan et al. analyze 12,099 incidents of testicular cancer in
the SEER database between 1973 and 1995. They report a steady increase
in the rate of testicular cancer in the US over this 23 yr time period:
from 3.61 per 100,000 to 5.44, a 51% increase. They also find that
the age of peak incidence has decreased and report this as a birth
cohort effect. This is particularly interesting because "a birth
cohort phenomenon occurs when a cohort, or study group, share a common
time period of birth and a similar risk of developing a particular
disease. Unlike most other cancers, testicular cancer, with a peak
incidence in the third decade of life, suggests a latency period that
involves some pre or postnatal stimulatory event that influences subsequent
tumor development. The widespread observation of a birth cohort correlation
for testicular cancer suggests that early or prolonged exposure to
some carcinogenic stimuli might be required for the subsequent development
of testicular cancer. "