JNCI J Natl Cancer Inst (2016) 108(10): djw110
doi: 10.1093/jnci/djw110
First published online July 10, 2016
Article
ARTICLE
Family History and Probability of Prostate Cancer,
Differentiated by Risk Category: A Nationwide
Population-Based Study
€ r Stattin
Ola Bratt, Linda Drevin, Olof Akre, Hans Garmo, Pa
Affiliations of authors: Department of Translational Medicine Urology, Division of Urological Cancers, Lund University, Sweden (OB); Department of Urology/CamPARI
€
Clinic, Cambridge University Hospitals, Cambridge, UK (OB);, Regional Cancer Centre, Uppsala/Orebro,
Uppsala University Hospital, Uppsala, Sweden (LD, HG);
Department of Urology, Karolinska Institute, Stockholm, Sweden (OA); Department of Surgical and Perioperative Sciences, Urology and Andrology, Umeå University,
Umeå, Sweden (PS); and Department of Surgical Sciences Uppsala University Hospital, Uppsala, Sweden (PS).
Correspondence to: Ola Bratt, MD, PhD, Associate Professor, Division of Urological Cancers, Department of Urology, Addenbrooke’s Hospital, Cambridge University
Hospitals NHS Foundation Trust, Hills Road, Cambridge CB2 0QQ, UK (e-mail: [email protected]).
Background: Familial prostate cancer risk estimates are inflated by clinically insignificant low-risk cancer, diagnosed after
prostate-specific antigen testing. We provide age-specific probabilities of non-low- and high-risk prostate cancer.
Methods: Fifty-one thousand, eight hundred ninety-seven brothers of 32 807 men with prostate cancer were identified in
Prostate Cancer data Base Sweden (PCBaSe). Nelson-Aalen estimates with 95% confidence intervals (CIs) were calculated
for cumulative, family history–stratified probabilities of any, non-low- (any of Gleason score 7, prostate-specific antigen
[PSA] 10 ng/mL, T3-4, N1, and/or M1) and high-risk prostate cancer (Gleason score 8 and/or T3-4 and/or PSA 20 ng/mL
and/or N1 and/or M1).
Results: The population probability of any prostate cancer was 4.8% (95% CI ¼ 4.8% to 4.9%) at age 65 years and 12.9% (95%
CI ¼ 12.8% to 12.9%) at age 75 years, of non-low-risk prostate cancer 2.8% (95% CI ¼ 2.7% to 2.8%) at age 65 years and 8.9% (95%
CI ¼ 8.8% to 8.9%) at age 75 years, and of high-risk prostate cancer 1.4% (95% CI ¼ 1.3% to 1.4%) at age 65 years and 5.2% (95%
CI ¼ 5.1% to 5.2%) at age 75 years. For men with one affected brother, probabilities of any prostate cancer were 14.9% (95%
CI ¼ 14.1% to 15.8%) at age 65 years and 30.3% (95% CI ¼ 29.3% to 31.3%) at age 75 years, of non-low-risk prostate cancer 7.3%
(95% CI ¼ 6.7% to 7.9%) at age 65 years and 18.8% (95% CI ¼ 17.9% to 19.6%) at age 75 years, and of high-risk prostate cancer
3.0% (95% CI ¼ 2.6% to 3.4%) at age 65 years and 8.9% (95% CI ¼ 8.2% to 9.5%) at age 75 years. Probabilities were higher for men
with a stronger family history. For example, men with two affected brothers had a 13.6% (95% CI ¼ 9.9% to 17.6 %) probability
of high-risk cancer at age 75 years.
Conclusions: The age-specific probabilities of non-low- and high-risk cancer presented here are more informative than relative risks of any prostate cancer and more suitable to use for counseling men with a family history of prostate cancer.
American and European clinical guidelines recommend men
with a family history of prostate cancer to obtain prostatespecific antigen (PSA) testing from age 40 to 50 years (1,2). The
recommendations are based on epidemiological studies showing a two- to five-fold increased relative risk of prostate cancer
for these men (3,4). However, in the present “PSA era,” familial
aggregation of prostate cancer is often caused by increased
diagnostic activity (PSA testing) among relatives of men with a
recently diagnosed prostate cancer, rather than by shared genetic predisposition (5). As a consequence, the magnitude of the
increased prostate cancer risk for men with a family history of
prostate cancer is to a large extent inflated by familial aggregation of “PSA detected,” clinically insignificant, low-risk prostate
cancer (5,6). The risk estimates presently used for counseling of
Received: December 7, 2015; Revised: February 11, 2016; Accepted: March 14, 2016
© The Author 2016. Published by Oxford University Press 2016.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/
licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
For commercial re-use, please contact [email protected]
1 of 7
ARTICLE
Abstract
2 of 7 | JNCI J Natl Cancer Inst, 2016, Vol. 108, No. 10
men with a family history of prostate cancer are therefore of
limited value. The clinical significance of a man’s probability of
high-risk, potentially lethal prostate cancer is certainly different
from that of his probability of any (possibly indolent) prostate
cancer. Moreover, current guidelines only include estimates of
relative risks of prostate cancer, despite recommendations that
information on medical risk to lay people should be conveyed
as absolute risks (7,8).
We have previously reported on the heritability of Gleason patterns (6). We now report on a nationwide, population-based register study aiming at providing age-specific probabilities of any,
non-low-risk, and high-risk prostate cancer for men with a family
history of prostate cancer and differentiating these probabilities
by the number of affected relatives, the relatives’ ages at diagnosis, and the risk category or severity of the relatives’ cancer.
NPCR 1998−2012
n = 129 389 men
with prostate cancer
Men born before 1932
n = 51 002
Men with no brothers*
n = 42 002
Methods
The Prostate Cancer Data Base Sweden
ARTICLE
The Prostate Cancer data Base Sweden (PCBaSe) 3.0 was created
through record linkages between the National Prostate Cancer
Register (NPCR) of Sweden and several other nationwide, population-based health care registers and demographic databases.
The NPCR of Sweden captures 98% of the prostate cancer cases
in the Swedish Cancer Registry, to which registration is mandated by law (9). PCBaSe has previously been described in detail
(10). The quality and completeness of the NPCR (11), as well as
of other Swedish national registers and databases, are high, and
notifications are regularly reviewed by Statistics Sweden.
Information in PCBaSe on family history was obtained from
the Multi-Generation Register, which is held by Statistics
Sweden. The register includes family information for all individuals born in Sweden since 1932 and who were still residents in
Sweden after in 1961 or later. In 2007, there were about nine million index persons in the register who could be linked to their
first-degree relatives (biological parents, siblings, and children).
Practically all first-degree relatives alive after 1990 are registered.
Figure 1 shows the selection of families for the final study cohort,
which included 51 897 brothers of 32 807 men with prostate cancer. Half-brothers of index cases with a common father were included (n ¼ 4035), whereas half-brothers with a common mother
were excluded. All brothers whose father was not registered in
the Swedish Cancer Registry as diagnosed with prostate cancer
were included in the analysis of the effect of the father’s prostate
cancer status, including those whose father was still alive, whose
father died young, or for whose father we had no information on
age at or cause of death (<10% of fathers). The PCBaSe project
was approved by the Research Ethics Board at Umeå University.
Definitions of Cancer Risk Categories
Prostate cancer diagnosed among brothers of index cases was
categorized as low risk (T1-2, Gleason score 6, PSA < 10 ng/
mL, Nx/N0, Mx/M0), non–low risk (any cancer not in the lowrisk category, ie, Gleason score 7 and/or T3-4 and/or PSA 10 ng/mL and/or N1 and/or M1), or high risk (any of Gleason
score 8, T3-4, PSA 20 ng/mL, N1 and/or M1). Most prostate
cancers among fathers of the index persons were diagnosed before 1998 and were thus not registered in the NPCR of Sweden.
Therefore, the severity of their cancers was categorized as low
risk, nonlethal, or lethal. Low risk was defined as prostate cancer diagnosed before the age of 75 years in men who lived at
Men who were not the
first among a set of
brothers to be diagnosed
with prostate cancer,
n = 3578
Included in the study:
Index cases: n = 32 807
Brothers of index cases:
n = 51 897
Figure 1. Flow chart of the selection process from Prostate Cancer data Base
Sweden. *Half-brothers of index cases with a common father were included
whereas half-brothers with a common mother were excluded. NPCR ¼ National
Prostate Cancer Register.
least 10 years after their diagnosis, nonlethal cancer as any
prostate cancer diagnosed in a father who did not die from prostate cancer (fathers who died of other causes within 10 years of
their prostate cancer diagnosis were included in this category),
and lethal cancer was any prostate cancer diagnosed in a father
who died from prostate cancer during the time of follow-up.
Statistical Analysis
Nelson-Aalen estimates were used to calculate cumulative
probabilities of prostate cancer with 95% confidence intervals
(CIs) for brothers of the index cases in PCBaSe and for men in
the general population, using age as time scale. The number of
men at risk (the denominator) in the estimates for the general
population was acquired from Statistics Sweden whereas the
denominator in the estimates for the brothers of the index cases
was acquired from PCBaSe. The number of prostate cancer
cases observed among men at risk (the numerator) was acquired from PCBaSe. The probability of prostate cancer was calculated as 1-exp(Nelson-Aalen estimator) (12). As there were
few men at risk before age 50 years, the precision of probability
estimates before that age would be low. We therefore present
probabilities conditioned on no prostate cancer diagnosed before age 50 years. The brothers of the index cases were followed
from the date of entry into one of the specific risk categories, as
defined by the dates for new prostate cancer diagnoses in their
family, to the date of their own diagnosis of prostate cancer,
O. Bratt et al. | 3 of 7
emigration, or death or to December 31, 2012, whichever occurred first. P values for comparisons between probabilities for
men with different family histories were calculated using the
Wald statistic. A P value of less than .05 was considered statistically significant, and all statistical tests were two-sided.
Results
For men in the general Swedish population, the probabilities of
any prostate cancer were 4.8% (95% CI ¼ 4.8% to 4.9%) at age 65
years, 12.9% (95% CI ¼ 12.8% to 12.9%) at age 75 years, and 17.1%
(95% CI ¼ 17.0% to 17.2%) at age 80 years; the probabilities of
non-low-risk prostate cancer were 2.8% (95% CI ¼ 2.7% to 2.8%)
at age 65 years, 8.9% (95% CI ¼ 8.8% to 8.9%) at age 75 years, and
12.7% (95% CI ¼ 12.6% to 12.8%) at age 80 years; and the probabilities of high-risk prostate cancer were 1.4% (95% CI ¼ 1.3% to
1.4%) at age 65 years, 5.2% (95% CI ¼ 5.1% to 5.2%) at age 75
years, and 8.2% (95% CI ¼ 8.1% to 8.2%) at age 80 years.
The probabilities of any, non-low-risk, and high-risk prostate
cancer increased with the number of affected family members
and with age. The probabilities at age 65 and 75 years are shown
in Table 1 and displayed graphically in Figures 2-5. For example,
for men with one affected brother, the probabilities of any prostate cancer were 14.9% (95% CI ¼ 14.1% to 15.8%) at age 65 years,
30.3% (95% CI ¼ 29.3% to 31.3%) at age 75 years, and 36.1% (95%
CI ¼ 34.8% to 37.4%) at age 80 years; the probabilities of non-lowrisk prostate cancer were 7.3% (95% CI ¼ 6.7% to 7.9%) at age 65
years, 18.8% (95% CI ¼ 17.9% to 19.6%) at age 75 years, and 24.2%
(95% CI ¼ 23.0% to 25.5%) at age 80 years; and the probabilities of
high-risk prostate cancer were 3.0% (95% CI ¼ 2.6% to 3.4%) at
age 65 years, 8.9% (95% CI ¼ 8.2% to 9.5%) at age 75 years, and
12.7% (95% CI ¼ 11.5% to 13.8%) at age 80 years. Men with two
brothers with prostate cancer had at age 75 years a 55.1% (95%
CI ¼ 49.8% to 59.9%) probability of any prostate cancer, a 33.2%
(95% CI ¼ 28.2% to 37.8%) probability of non-low-risk cancer, and
a 13.6% (95% CI ¼ 9.9% to 17.6%) probability of high-risk cancer.
Table 1. Probabilities (95% confidence intervals) of prostate cancer at age 65 and 75 years in Swedish men according to their family history of
prostate cancer*
Family history
†
Population risk
1 brother, any PCa
1 brother low-risk PCa
1 brother non-low-risk PCa
1 brother high-risk PCa
Father (any age) þ brother PCa
Father (75 y) þ brother
PCa
Father (<75 y) þ brother PCa
Father low-risk PCa þ brother
PCa
Father low-risk PCa þ brother
low-risk PCa
Father low-risk PCa þ brother
non-low-risk PCa
Father low-risk PCa þ brother
high-risk PCa
Father nonlethal PCa þ brother
PCa
Father nonlethal PCa þ brother
low-risk PCa
Father nonlethal PCa þ brother
non-low-risk PCa
Father nonlethal PCa þ brother
high-risk PCa
Father died of PCa þ brother
PCa
Father died of PCa þ brother
low-risk PCa
Father died of PCa þ brother
non-low-risk PCa
Father died of PCa þ brother
high-risk PCa
2 brothers PCa
Father þ 2 brothers PCa
Non-low-risk PCa, %
High-risk PCa, %
No.
By age 65 y
By age 75 y
By age 65 y
By age 75 y
By age 65 y
By age 75 y
NA
38 921
13 660
24 404
12 769
7757
3894
4.8 (4.8 to 4.9)
14.9 (14.1 to 15.8)
13.8 (12.5 to 15.1)
15.7 (14.5 to 16.9)
16.1 (14.5 to 17.6)
29.8 (27.0 to 32.5)
26.5 (22.1 to 30.6)
12.9 (12.8 to 12.9)
30.3 (29.3 to 31.3)
28.8 (27.1 to 30.4)
31.4 (30.0 to 32.7)
31.7 (29.9 to 33.3)
47.8 (45.1 to 50.3)
45.2 (41.2 to 48.9)
2.8 (2.7 to 2.8)
7.3 (6.7 to 7.9)
6.3 (5.4 to 7.1)
7.9 (7.1 to 8.7)
8.0 (6.9 to 9.0)
13.7 (11.8 to 15.5)
11.1 (8.4 to 13.6)
8.9 (8.8 to 8.9)
18.8 (17.9 to 19.6)
16.9 (15.5 to 18.2)
19.9 (18.8 to 21.0)
19.7 (18.3 to 21.1)
28.2 (25.8 to 30.5)
26.2 (22.9 to 29.4)
1.4 (1.3 to 1.4)
3.0 (2.6 to 3.4)
2.4 (1.8 to 3.0)
3.4 (2.8 to 3.9)
3.4 (2.6 to 4.1)
5.6 (4.4 to 6.7)
4.5 (2.9 to 6.1)
5.2 (5.1 to 5.2)
8.9 (8.2 to 9.5)
8.0 (7.0 to 9.1)
9.4 (8.5 to 10.2)
9.3 (8.2 to 10.4)
13.8 (11.9 to 15.6)
12.7 (10.1 to 15.2)
3863 33.0 (29.3 to 36.5) 50.1 (46.5 to 53.6) 16.3 (13.6 to 18.9) 30.0 (26.5 to 33.2) 6.9 (5.1 to 8.6)
3007 28.9 (19.3 to 37.4) 47.3 (39.7 to 54.0) 11.0 (7.2 to 14.7) 26.2 (22.0 to 30.1) 3.9 (2.5 to 5.3)
15.0 (12.2 to 17.8)
12.2 (9.7 to 14.7)
1175 21.0 (15.4 to 26.2) 41.5 (35.3 to 47.1) 9.4 (5.8 to 12.8) 24.5 (19.0 to 29.6) 3.6 (1.3 to 5.8)
10.3 (6.3 to 14.1)
1770 34.4 (19.3 to 46.7) 51.5 (39.9 to 60.9) 12.7 (6.3 to 18.6) 27.8 (21.5 to 33.6) 4.2 (2.4 to 6.0)
13.2 (10.0 to 16.4)
926 36.2 (16.1 to 51.5) 50.6 (34.5 to 62.7) 13.4 (3.8 to 22.0) 25.2 (15.9 to 33.4) 2.6 (0.7 to 4.4)
10.6 (6.6 to 14.5)
6379 29.0 (25.6 to 32.3) 47.4 (44.3 to 50.3) 13.1 (10.9 to 15.2) 27.7 (25.1 to 30.3) 4.9 (3.8 to 6.0)
13.3 (11.4 to 15.2)
2579 26.4 (20.9 to 31.5) 44.9 (39.8 to 49.6) 11.9 (8.8 to 15.0) 25.6 (21.5 to 29.5) 4.2 (2.5 to 5.8)
10.8 (7.9 to 13.6)
3671 31.1 (26.6 to 35.4) 49.2 (45.2 to 52.9) 14.0 (11.0 to 16.9) 29.0 (25.5 to 32.4) 5.6 (4.0 to 7.1)
14.7 (12.1 to 17.2)
1841 34.2 (27.1 to 40.6) 50.4 (44.3 to 55.8) 14.4 (9.9 to 18.6) 28.2 (23.3 to 32.9) 5.1 (3.1 to 7.0)
14.2 (10.7 to 17.5)
1183 29.8 (21.0 to 37.6) 45.0 (36.3 to 52.6) 15.3 (9.9 to 20.4) 27.5 (20.3 to 34.0) 6.6 (2.6 to 10.5) 12.7 (6.6 to 18.4)
519 33.6 (18.8 to 45.7) 51.9 (36.7 to 63.5) 14.1 (5.4 to 22.1) 31.3 (16.7 to 43.3) 5.3 (-0.5 to 10.8) 15.1 (1.2 to 27.1)
646 25.1 (17.4 to 32.2) 39.7 (30.4 to 47.7) 16.4 (9.3 to 22.8) 26.4 (17.6 to 34.3) 7.9 (2.1 to 13.3) 12.6 (5.1 to 19.4)
292 22.4 (9.7 to 33.3) 36.0 (22.2 to 47.3) 18.0 (5.5 to 28.8) 24.7 (11.7 to 35.9) 12.0 (0.5 to 22.3) 13.4 (1.7 to 23.7)
2343 34.4 (28.1 to 40.1) 55.1 (49.8 to 59.9) 14.3 (10.0 to 18.4) 33.2 (28.2 to 37.8) 2.7 (0.8 to 4.5) 13.6 (9.9 to 17.1)
709 43.9 (33.7 to 52.5) 63.6 (54.9 to 70.6) 23.7 (14.0 to 32.2) 39.4 (29.2 to 48.1) 11.4 (3.6 to 18.5) 20.5 (11.3 to 28.8)
*The brothers’ cancers were categorized as low risk (T1 to 2, Gleason score 6, prostate-specific antigen [PSA] < 10 ng/mL, Nx/N0, Mx/M0), non-low-risk (Gleason score
7 and/or T3 to 4 and/or PSA 10 ng/mL and/or N1 and/or M1) or high risk (any of Gleason score 8, T3 to 4, PSA 20 ng/mL, N1 and/or M1). The severity of the fathers’ cancer was categorized as low risk: diagnosed before age 75 years in men who lived at least 10 years after diagnosis; nonlethal cancer: diagnosed in men who did
not die from prostate cancer; or lethal cancer: diagnosed in men who died from prostate cancer during follow-up. PCa ¼ prostate cancer.
†
The risk for men with a negative family history is approximately 10% lower than the population risk.
ARTICLE
Any PCa, %
0.8
4 of 7 | JNCI J Natl Cancer Inst, 2016, Vol. 108, No. 10
Family history of prostate cancer
Father and two brothers
Two brothers
Father and one brother
One brother
0.4
0.0
0.2
Probability of prostate cancer
0.6
Population
50
55
60
65
70
75
80
Age, y
Figure 2. Probability of any prostate cancer according to the number of affected family members, given the conditional probability of no prostate cancer at age 50 years.
Family history of prostate cancer
50
60
70
Age, y
80
50
60
70
80
0.8
0.8
0.4
0.6
0.6
0.4
0.0
0.0
0.2
0.2
0.4
0.2
0.0
ARTICLE
0.0
0.2
0.2
0.4
0.4
0.6
0.6
Any
Non-low-risk
High-risk
Father and
two brothers
Two brothers
0.8
0.8
Type of prostate cancer
0.6
Father and
one brother
One brother
0.0
Probability of prostate cancer
0.8
Population
50
Age, y
60
70
80
Age, y
50
60
70
Age, y
80
50
60
70
80
Age, y
Figure 3. Probabilities of any, non-low-risk and high-risk (including metastatic) prostate cancer according to the number of affected family members, given the conditional probability of no prostate cancer at age 50 years.
Most men with a family history of prostate cancer had a 30%
to 60% probability of any prostate cancer at age 75 years, but
around half of them were diagnosed with low-risk cancer. Their
probability of high-risk prostate cancer was typically one-sixth
to one-fourth of the probability of any prostate cancer (Table 1).
The relative risks of familial prostate cancer decreased with
age, for example, of any prostate cancer for brothers of men
with prostate cancer from 3.1 (95% CI ¼ 2.9 to 3.3) at age 65 years
to 2.4 (95% CI ¼ 2.3 to 2.4) at age 75 years and to 2.1 (95% CI ¼ 2.0
to 2.2) age 80 years.
For men with a father diagnosed before age 75 years, the
probabilities of non-low-risk and of high-risk cancer at age 65
years were higher than for men with a father diagnosed after
age 75 years (P ¼ .005 and P ¼ .05) (Table 1 and Figure 4). We did
not analyze the effect of the age at diagnosis of an affected
brother because the age at diagnosis for affected brothers
covaries as they are likely to be of similar age. For example,
brothers of men diagnosed in their fifties are younger and
therefore at risk of being diagnosed in their fifties or sixties
whereas most brothers of men diagnosed with prostate cancer
at age 75 years are in their seventies (had these brothers been
diagnosed in their fifties or sixties, they would have been the index case the family). Neither the brothers’ nor the fathers’ cancer’s risk category statistically significantly affected the
probability of any, non-low- or high-risk disease (Table 1).
Discussion
We present, for the first time, age-specific probabilities of nonlow- and high-risk prostate cancer for men with a family history
of prostate cancer. Most of these men had a 30% to 60%
O. Bratt et al. | 5 of 7
0.7
0.6
< 75
≥ 75
0.6
High-risk prostate cancer
0.6
Age, y
0.5
0.4
0.3
0.2
0.0
0.0
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
0.4
0.5
0.5
Population
0.1
Probability of prostate cancer
Non-low-risk prostate cancer
0.7
0.7
Any prostate cancer
50
55
60
65
70
75
80
50
55
Age, y
60
65
Age, y
70
75
80
50
55
60
65
70
75
80
Age, y
Figure 4. Probability of any prostate cancer according to the father’s age at diagnosis in men with a father and a brother with prostate cancer, given the conditional
probability of any prostate cancer at age 75 years, but around
half of the diagnosed cancers were in the low-risk category and
of little clinical significance. The men’s probability of high-risk
prostate cancer was much lower, typically one-sixth to onefourth of the probability of any prostate cancer. For example,
men with two brothers with prostate cancer had at age 75 years
a 55.1% probability of any prostate cancer, a 33.2% probability of
non-low-risk cancer, and a 14.6% probability of high-risk cancer.
Clearly, the probabilities of non-low-risk and, in particular,
high-risk cancer are highly relevant when counseling men with
familial prostate cancer. We suggest that these probabilities are
included in future clinical guidelines.
In agreement with the results from our previous studies on
familial prostate cancer (5,6), the present study indicates that
PSA testing leads to an “inflation” of familial prostate cancer
risks. Only 2.0% of men in the general population were diagnosed with low-risk cancer at age 65 years, compared with 7.6%
of men with one affected brother and 20.1% of men with two affected brothers. These absolute differences are much larger
than the corresponding differences for high-risk cancer.
Because low-risk prostate cancer almost always is diagnosed after PSA testing, our results strongly suggest that PSA testing is
more common among men with than in men without a family
history of prostate cancer.
Almost all previous studies of familial prostate cancer have
reported relative risk estimates (relative risks, rate ratios, hazard ratios, or standard incidence ratios). The notion of a relative
risk may be difficult to grasp, and absolute risks are therefore
recommended for clinical counseling (7,8). The age-specific
probabilities provided by our study are in essence absolute risks
and will be useful for clinicians who discuss the risk of familial
prostate cancer with their patients. Experts in risk communication recommend visualization of absolute risks in the form of
pictograms (7,8). We therefore provide two pairs of pictograms
as examples of how our results may be presented to laypeople
(Figure 5).
The probability of high-risk prostate cancer was affected by
the number of affected family members and by the severity of
the prostate cancer diagnosed in a father, but not by the brother’s cancer’s risk category: The probability of a future high-risk
prostate cancer for brothers of men with a low-risk cancer was
only slightly lower than for brothers of men with a high-risk
cancer. This is a clinically important finding; brothers of men
diagnosed with a low-risk prostate cancer should not be recommended against PSA testing just because their brother’s cancer
might have been considered clinically insignificant.
We have previously shown that brothers of men with
Gleason score 8-10 cancer are at a particularly high relative risk
of cancer with the same Gleason score (6). The proportion of
brother pairs concordant for Gleason score 8-10 cancer in that
study was, however, small. This means that the effect size of
Gleason score 8-10 cancer heritability on the absolute probability reported in the present study is small and compatible with
the only slightly higher probability of high-risk cancer among
brothers of men with high-risk cancer.
Our study shows that relative risks of familial prostate cancer decrease with age. The relative risk of any prostate cancer
for brothers of men with prostate cancer was 3.1 at age 65 years,
which is exactly the same as reported in a meta-analysis of 16
studies (3), but decreased to 2.4 at age 75 years and to 2.1 at age
80 years. This is in agreement with a previous study that reported absolute risks of prostate cancer for men with and without a family history of the disease (13). Therefore, relative risk
estimates derived from studies on younger men should not be
multiplied with the lifetime risk in the population to achieve an
estimate of the absolute lifetime risk.
Strengths of our study include the use of nationwide,
population-based, high-quality register data with almost complete capture rates (10,11). The statistical precision of our calculations was therefore high. The narrow confidence intervals
should not, however, be confused with high external validity.
The incidences of any and high-risk prostate cancer vary geographically; the incidence of low-risk prostate cancer also
changes over time because of varying diagnostic activity (PSA
testing). The reported probabilities of any prostate cancer
should be regarded as approximations of the probabilities for
men only in populations with a high prostate cancer incidence
and a high uptake of PSA testing, such as in North America,
Australia, New Zealand, and north-western Europe. The incidence of high-risk prostate cancer is not much affected by PSA
testing, so the reported probabilities of high-risk disease are applicable also for men in populations with a high incidence and a
low uptake of PSA testing, such as in the Caribbean and in some
parts of Africa. Moreover, despite the inclusion of more than
50000 men at risk in total, the number of men at risk before age
55 years was low in many of the subgroups. The confidence
ARTICLE
probability of no prostate cancer at age 50 years.
6 of 7 | JNCI J Natl Cancer Inst, 2016, Vol. 108, No. 10
ARTICLE
Figure 5. Pictograms of the probabilities of no (white men), low-risk (light gray men), intermediate risk (dark gray men), and high-risk, including metastatic (black
men), prostate cancer. A) Average population probabilities at age 65 years. B) Probabilities at age 65 years for men with a father and one brother with prostate cancer. C)
Average population probabilities at age 75 years. D) Probabilities at age 75 years for men with a father and one brother with prostate cancer.
O. Bratt et al. | 7 of 7
Funding
This work was supported by the Swedish Research Council
(PCBaSe: grant number 825-2012-5047) and by the Swedish
Cancer Foundation (Ola Bratt: grant number 2012/475).
Notes
The study funders had no role in the design of the study; the
collection, analysis, or interpretation of the data; the writing of
the manuscript; or the decision to submit the manuscript for
publication.
The Prostate Cancer data Base Sweden was made possible
by the continuous work of the National Prostate Cancer Register
€ r Stattin (chairman), Anders
of Sweden steering group: Pa
Widmark, Camilla Thellenberg, Ove Andrén, Anna Bill-Axelson,
Ann-Sofi Fransson, Magnus Törnblom, Stefan Carlsson, Marie
€ lm-Eriksson, Bill Pettersson, David Robinson, Mats Andén,
Hja
Jan-Erik Damber, Jonas Hugosson, Ingela Franck-Lissbrant,
Maria Nyberg, Göran Ahlgren, Ola Bratt, René Blom, Lars
Egevad, Calle Waller, Olof Akre, Per Fransson, Eva Johansson,
Fredrik Sandin, Karin Hellström.
References
1. Heidenreich A, Bellmunt J, Briers E, et al. European Association of Urology
Guidelines on Prostate Cancer. 2015. http://uroweb.org/wp-content/uploads/
EAU-Guidelines-Prostate-Cancer-2015-v2.pdf. Accessed December 2, 2015.
2. Carter HB, Albertsen PC, Barry MJ, et al. Early detection of prostate cancer:
AUA Guideline. J Urol. 2013;190(2):419–426.
3. Kicinski M, Vangronsveld J, Nawrot TS. An epidemiological reappraisal of the
familial aggregation of prostate cancer: a meta-analysis. PloS One. 2011;6(10):
e27130.
4. Albright F, Stephenson RA, Agarwal N, et al. Prostate cancer risk prediction
based on complete prostate cancer family history. Prostate. 2015;75(4):
390–398.
5. Bratt O, Garmo H, Adolfsson J, et al. Effects of prostate-specific antigen testing on familial prostate cancer risk estimates. J Natl Cancer Inst. 2010;102(17):
1336–1343.
6. Jansson KF, Akre O, Garmo H, et al. Concordance of tumor differentiation
among brothers with prostate cancer. Eur Urol. 2012;62(4):656–661.
7. Zipkin DA, Umscheid CA, Keating NL, et al. Evidence-based risk communication: a systematic review. Ann Int Med. 2014;161(4):270–280.
8. Fagerlin A, Zikmund-Fisher BJ, Ubel PA. Helping patients decide: ten steps to
better risk communication. J Natl Cancer Inst. 2011;103(19):1436–1443.
9. Tomic K, Berglund A, Robinson D, et al. Capture rate and representativity of
The National Prostate Cancer Register of Sweden. Acta Oncol. 2014:1–6.
10. Van Hemelrijck M, Wigertz A, Sandin F, et al. Cohort Profile: the National
Prostate Cancer Register of Sweden and Prostate Cancer data Base Sweden
2.0. Int J Epidemiol. 2013;42(4):956–967.
11. Tomic K, Sandin F, Wigertz A, Robinson D, Lambe M, Stattin P. Evaluation of
data quality in the National Prostate Cancer Register of Sweden. Eur J Cancer.
2015;51(1):101–111.
12. Breslow NE. Discussion of the paper by D. R. Cox. Statist Soc B. 1972;34:
216–217.
13. Grönberg H, Wiklund F, Damber JE. Age specific risks of familial prostate carcinoma: a basis for screening recommendations in high risk populations.
Cancer. 1999;86(3):477–483.
ARTICLE
intervals of the probabilities shown in the figures for men younger than 55 years are therefore wider than the ones shown for
men at age 65 years in Table 1.
Another limitation is that the sons of men registered in
PCBaSe were too young to allow analysis of their cancer risk;
only probabilities for the index men’s brothers were analyzed.
The risk of prostate cancer for men whose father is the only affected family member is lower than the risk for men with an affected brother (3). A meta-analysis reported risk ratios of 2.4 for
sons and of 3.1 for brothers of men with prostate cancer (3).
We had detailed, complete information on cancer characteristics at the time of diagnosis for the index cases’ brothers. No
such detailed information was available to categorize the severity of the fathers’ cancers. The ad hoc nature of the criteria for
these categories may have contributed to the lack of association
between the fathers’ and their sons’ cancer risk categories.
Moreover, the categorization of the brothers’ cancers was based
on the diagnostic prostate biopsies only. As some men diagnosed with low-risk cancer have an undetected prostate cancer
of higher grade, the tendency for aggressive prostate cancer to
run in families may be stronger than suggested by our results.
To summarize, we used nationwide, population-based registers
to calculate age-specific probabilities of any, non-low-risk, and
high-risk prostate cancer for men with and without a known family history of prostate cancer. The probabilities of high-risk prostate
cancer were substantially lower than of any prostate cancer. As
the probability of any prostate cancer in populations with widespread PSA testing is inflated by clinically insignificant cancer, the
probabilities of non-low- and high-risk prostate cancer are more
appropriate for counseling men with a family history of prostate
cancer. We suggest that age-specific probabilities of non-low- and
high-risk cancer are incorporated in future clinical guidelines.