Schizophrenia Research 153 (2014) 214–219

Contents lists available at ScienceDirect

Schizophrenia Research
journal homepage: www.elsevier.com/locate/schres

Population impact of familial and environmental risk factors for

schizophrenia: A nationwide study
Holger J. Sørensen a,d,⁎, Philip R. Nielsen b,d, Carsten B. Pedersen b,c,d, Michael E. Benros a,b,d,
Merete Nordentoft a,d, Preben B. Mortensen b,c,d
a
Psychiatric Centre Copenhagen, Capital Region of Denmark, University Hospital of Copenhagen, Bispebjerg Bakke 23, 2400 NV, Denmark
b
National Centre for Register-Based Research, Aarhus University, Business and Social Sciences, Aarhus, Denmark
c
Centre for Integrated Register-based Research, CIRRAU, Aarhus University, Aarhus, Denmark
d
The Lundbeck Foundation Initiative for Integrative Psychiatric Research, iPSYCH, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: Although several studies have examined the relative contributions of familial and environmental risk factors for
Received 30 September 2013 schizophrenia, few have additionally examined the predictive power on the individual level and simultaneously
Received in revised form 3 December 2013 examined the population impact associated with a wide range of familial and environmental risk factors. The
Accepted 1 January 2014
authors present rate ratios (IRR), population-attributable risks (PAR) and sex-specific cumulative incidences of
Available online 17 January 2014
the following risk factors: parental history of mental illness, urban place of birth, advanced paternal age, parental
Keywords:
loss and immigration status. We established a population-based cohort of 2,486,646 million persons born in
Schizophrenia Denmark between 1 January 1955 and 31 December 1993 using Danish registers. We found that PAR associated
Psychiatric family history with urban birth was 11.73%; PAR associated with one, respectively 2, parent(s) with schizophrenia was 2.67%
Urban birth and 0.12%. PAR associated with second-generation immigration was 0.70%. Highest cumulative incidence
Environmental risk factors (CI = 20.23%; 95% CI = 18.10–22.62) was found in male offspring of 2 parents with schizophrenia. Cumulative
Register-based incidences for male offspring or female offspring of a parent with schizophrenia were 9.53% (95% CI = 7.71–
11.79), and 4.89%, (95% CI 4.50–5.31). The study showed that risk factors with highest predictive power on the
individual level have a relatively low population impact. The challenge in future studies with direct genetic
data is to examine gene-environmental interactions that can move research beyond current approaches and
seek to achieve higher predictive power on the individual level and higher population impact.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction Genetic and environmental factors are associated with psychosis

Evidence from family, twin and adoptive studies suggest that genetic
transmission accounts for most of the familial aggregation of schizo-
phrenia (Kendler and Diehl, 1993). Registry-based epidemiologic
research supports that the risk of schizophrenia is associated not only
with a family history of schizophrenia (Gottesman et al., 2010) but
also with other categories of mental disorders in first-degree relatives
(Dean et al., 2010; Mortensen et al., 2010). In addition to a family history
of mental illness, large-scale epidemiologic studies have identified a
wide range of risk factors (for a review see Matheson et al. (2011)),
including urban birth (Pedersen and Mortensen, 2001; Pedersen,
2006; Pedersen and Mortensen, 2006), advancing paternal age (El-
Saadi et al., 2004; Miller et al., 2011; Petersen et al., 2011), migrant back-
ground (Weiser et al., 2008; Cantor-Graae and Pedersen, 2013), and
parental loss (Granville-Grossman, 1966; Laursen et al., 2007).
⁎ Corresponding author at: Mental Health Centre Copenhagen, Bispebjerg Bakke 23, DK
2400 NV, Copenhagen, Denmark. Tel.: +45 38647441; fax: +45 3864.
E-mail address: holger.jelling.soerensen@regionh.dk (H.J. Sørensen).
risk, but the latter present more tangible markers for prevention
(Kirkbride et al., 2010). In this context, relatively little attention has
been given to the question of the magnitude of the population impact
associated with a wide range of (genetic and environmental) risk
factors for schizophrenia. Reporting on the power to predict risk on an
individual level is important for preventive research as is the knowledge
of the feasibility of individual screening and intervention in relation to
severe mental illness (Kirkbride and Jones, 2011). While the importance
of fully elucidating the likely multifactorial, multilevel, polygenetic and
eco-epidemiologic basis of schizophrenia needs to be stressed, it is of
added importance to provide empirical data on epidemiological asso-
ciations from large-scale comprehensive studies. These continued
research efforts may in turn provide important new clues to the risk ar-
chitecture of schizophrenia. In this context, epidemiologic measures
such as the population attributable risk (PAR) and cumulative inci-
dences (CI) are important to the researcher, for instance when consider-
ing whether risk factors with high predictive power on the individual
level have high, modest or low population impact. In an attempt to
document on the population impact of familial and environmental risk
factors for schizophrenia, we initiated a register-based investigation in

0920-9964/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.schres.2014.01.008
 H.J. Sørensen et al. / Schizophrenia Research 153 (2014) 214–219
a single large sample in order to examine familial and environmental
risk factors for schizophrenia jointly. Furthermore, we examined sex-
specific lifetime absolute risks of these factors, as well as the relative
risks (IRR) and the population attributable risks (PAR). The following
risk factors could be jointly examined in our study: parental history of
mental illness, urban place of birth, advanced paternal age, parental
loss and immigration status.
2. Methods
2.1. Study population
We used data from the Danish Civil Registration System (CRS) to ob-
tain information on Danish people and their parents. The register was
established in 1968, where all people alive and living in Denmark
were registered. Among many other variables, it includes information
on personal identification number, sex, date of birth, continuously up-
dated information on vital status, and personal identification number
of parents. The personal identification number is used in all national
registers enabling accurate linkage between registers. Our study
population included all persons born in Denmark between 1 January
1955 and 31 December 1993, who were alive at their 15th birthday,
and where information on maternal identity was available (2,486,646
million people).
2.2. Assessment of schizophrenia and mental illness in parents
The study population and their mothers and fathers were linked
with the Danish Psychiatric Central Register (Munk-Jorgensen and
Mortensen, 1997), which was computerized in 1969. The Danish Psy-
chiatric Central Register contains data on all admissions to Danish psy-
chiatric in-patient facilities, and, from 1995, information on outpatient
visits to psychiatric departments was included in the register. At present
the register includes data on approximately 720,000 persons and 3.2
million contacts. From 1969 to 1993, the diagnostic system used was
the Danish modification of International Classification of Diseases, 8th
revision (ICD-8) (WHO, 1967), and, from 1994, the diagnostic system
used was the International Classification of Diseases, 10th revision
(ICD-10) (WHO, 1992). Cohort members were classified as cases if
they had been admitted to a psychiatric hospital or had been in outpa-
tient care with a diagnosis of schizophrenia (ICD-8 code 295 or ICD-10
code F20). Date of onset was defined as first day of first contact (in- or
outpatient) with a diagnosis of schizophrenia. Parents were categorized
hierarchically with a history of schizophrenia, schizophrenia-like psy-
choses, or other mental disorders (any ICD-8 or ICD-10 diagnosis), re-
spectively, if they had been admitted to a psychiatric hospital or in
outpatient care with one of these diagnoses. If both parents suffered
from different disorders, only the disorder with the highest hierarchy
was recorded.
2.3. Assessment of exposures
Municipalities in Denmark were classified according to degree of
urbanization by Statistics Denmark (Pedersen et al., 2006): capital,
capital suburb, provincial city, provincial town, or rural areas, which
have been used in previous studies. Maternal and paternal age at birth
was subdivided into 5 and 6 categories, respectively, as in the previous
studies. Information on parental loss was obtained from the Danish Civil
Registration System. Parental origin was defined as follows: second-
generation immigrants by both parents (mother and father born
abroad), second-generation immigrants by mother (mother born
abroad), second-generation immigrants by father (father born abroad),
native Danes (both parents born in Denmark), as previously (Cantor-
Graae and Pedersen, 2007).
215
2.4. Statistical analyses
A total of 2,486,646 people born in Denmark 1955–1993 were
followed from their 15th birthday until onset of schizophrenia, death,
emigration from Denmark, or 30 June 2009, whichever came first. The
relative risk of schizophrenia was estimated by Poisson regression. All
relative risks were adjusted for calendar year, age and its interaction
with sex. In addition we adjusted for a slight change in the age and
sex-specific incidence during the study period. As Cox regression is
very computer intensive for large studies, we used Poisson regression
as an approximation.
Age, calendar year, history of mental illness in parents and parental
loss were treated as time-dependent variables, whereas all other
variables were considered time independent. Age was categorized in
1-year intervals from age 15–19, in two-year intervals from 20 to 29,
and in 5-year intervals thereafter. Calendar years were categorized as
1970–1972 and in 1-year periods thereafter. P values were based on
likelihood ratio tests.
The cumulative incidences were calculated as the percentages of
persons in the population who had developed schizophrenia before a
given age, taking into account that people may die or migrate before
the onset of schizophrenia. When calculating the cumulative incidences,
history of mental illness in parents and parental loss were treated as
fixed variables evaluated at the time of the 15th birthday. These
analyses were made for each sex.
The population-attributable risk (the fraction of the total number
of cases in the population that would not have occurred if all people
had the same incidence rate as persons in the reference group) was
estimated as described by Bruzzi et al. (1985).
3. Results
3.1. Incidence rates, relative risks and population-attributable risk
Among the 2,486,646 people born in Denmark between the years
1955–1993, a total of 17,389 were registered with schizophrenia during
the 50,281,105 person-years of follow-up (overall incidence rate of 3.46
per 10,000 person-years at risk). There were 3420 persons with schizo-
phrenia and a parental history of mental illness, 548 with schizophrenia
and one parent with schizophrenia, and 5 persons with schizophrenia
where both parents also had received the same diagnosis. A total of
1558 persons with schizophrenia had second-generation immigration
background. Table 1 shows the crude and adjusted IRR's and PAR's
according to parental history of mental illness, place of birth, maternal
age, paternal age, maternal loss, paternal loss and immigration status.
Estimates in the first adjustment were adjusted for calendar year, age
and its interaction with sex, whereas estimates in the second adjust-
ment were also adjusted for all other variables in the table. We found
that the effects of most risk factors were attenuated somewhat after
mutual adjustment, but all risk factors remained highly significant. In
the following we only refer to the second adjustment.
3.2. Urbanization at birth
Our results suggest that the higher the degree of urbanization at
place of birth is, the higher the relative risk of schizophrenia. Using
birth in a rural area as the reference, people born in the capital had a
1.80-fold increased risk of schizophrenia, and the PAR associated with
birth in the capital was 11.73%, i.e. if persons born in the capital had
the same risk as persons born in the rural area, then 11.73% of the
cases in the total Danish population would not have occurred. The
highest PAR for any of the risk factors examined in this analysis was
for urban birth (11.73%). The cumulative risk (i.e. sex-specific lifetime
risk of schizophrenia) relative to the degree of urbanization of birth is
shown in Table 2. The effects of degree of urbanization are in the same
direction for men and women, and the highest lifetime risk is 2.16%
216 H.J. Sørensen et al. / Schizophrenia Research 153 (2014) 214–219

Table 1
Incidence rate ratios for schizophrenia by a family history, maternal age and paternal age, among 2,486,646 people born in Denmark 1955–1993, of which 17,389 developed schizophrenia.

Cases of Incidence Incidence rate ratios first adjustment Incidence rate ratios second adjustment Population attributable
schizophrenia ratea (95% confidence interval)b (95% confidence interval)c risk % (PAR)d

Overall 17,389 3.5

Family history of mental illnesse (p = b.00001) (p = b.00001)
Both parents schizophrenic 21 86.36 25.94 (16.36–38.70) 20.96 (13.22–31.28) 0.12
One parent schizophrenic 548 23.47 7.58 (6.95–8.26) 6.60 (6.04–7.19) 2.67
Both parents schizophrenia-like psychosis 5 22.71 7.47 (2.68–16.05) 6.13 (2.20–13.17) 0.02
One parent schizophrenia-like Psychosis 599 12.87 4.41 (4.06–4.78) 4.00 (3.68–4.34) 2.58
Both parents psychiatric Contact 467 11.44 3.99 (3.64–4.38) 3.49 (3.17–3.82) 1.92
One parent psychiatric contact 3420 6.41 2.23 (2.15–2.32) 2.07 (2.00–2.16) 10.19
Parents with no psychiatric contact 12,329 2.81 1.00 (reference) 1.00 (reference)
Urbanicity (p = b.00001) (p = b.00001)
Capital 4600 5.2 2.13 (2.04–2.23) 1.80 (1.72–1.88) 11.73
Capital suburb 1902 4.1 1.53 (1.45–1.62) 1.40 (1.32–1.48) 3.12
Provincial city 2174 3.4 1.33 (1.26–1.41) 1.28 (1.21–1.35) 2.71
Provincial town 5245 3.0 1.18 (1.13–1.23) 1.14 (1.09–1.19) 3.59
Rural area 3468 2.6 1.00 (reference) 1.00 (reference)
Maternal loss (p = b.00001) (p = b.00001)
Mother death 1648 4.3 1.48 (1.40–1.56) 1.20 (1.14–1.27) 1.56
Mother alive 15,741 3.4 1.00 (reference) 1.00 (reference)
Paternal losse (p = b.00001) (p = b.00001)
Father death 2953 4.0 1.43 (1.37–1.49) 1.16 (1.11–1.21) 2.34
Father alive 13,715 3.3 1.00 (reference) 1.00 (reference)
Immigratione (p = b.00001) (p = b.00001)
2nd generation immigrants, both 325 8.7 2.05 (1.83–2.29) 1.59 (1.42–1.77) 0.70
2nd generation, mother 643 5.8 1.70 (1.57–1.83) 1.47 (1.36–1.58) 1.23
2nd generation, father 590 6.5 1.82 (1.67–1.97) 1.49 (1.37–1.62) 1.12
Parents native Danes 14,934 3.3 1.00 (reference) 1.00 (reference)
Maternal age (p = b.00001) (p = b.00001)
12–19 1674 4.0 1.32 (1.25–1.39) 1.19 (1.12–1.27) 1.56
20–24 5833 3.4 1.08 (1.04–1.12) 1.07 (1.03–1.12) 2.20
25–29 5365 3.3 1.00 (reference) 1.00 (reference)
30–34 2939 3.5 1.07 (1.02–1.12) 0.99 (0.95–1.04) −0.11
≥35 1578 3.8 1.23 (1.16–1.30) 1.03 (0.96–1.10) 0.28
Paternal agee (p = b.00001) (p = b.00001)
12–19 376 3.9 1.33 (1.20–1.47) 1.03 (0.92–1.15) 0.07
20–24 3441 3.5 1.13 (1.08–1.18) 1.02 (0.97–1.07) 0.41
25–29 5326 3.2 1.00 (reference) 1.00 (reference)
30–34 3916 3.3 1.04 (1.00–1.08) 1.06 (1.02–1.11) 1.32
35–44 3039 3.6 1.18 (1.13–1.23) 1.17 (1.10–1.23) 2.48
≥45 570 4.4 1.51 (1.38–1.65) 1.38 (1.25–1.51) 0.89

(44,219 people, 712 cases, 1.2 million person years).

a
The incidence rate measures the number of new cases per 10,000 person years at risk.
b
Adjusted for calendar year, age and its interaction with gender.
c
Same as first adjustment and mutually adjusted for all variables in the table.
d
The population attributable risk measures the fraction of the total number of cases in the population that would not have occurred if persons in each exposure group had the same
incidence rate as persons in the reference group. The estimates shown are based on the second adjustment.
e
People with missing father were not shown in the total and were treated as a separate category in the analysis.

for males born in capital which is more than twice the estimate for
males born in rural areas.
3.3. Family history
Using offspring of parents with no psychiatric contact as the refer-
ence, all categories of parental psychiatric diagnoses had significantly
increased relative risks. The incidence rate in offspring whose both
parents had been diagnosed with schizophrenia was 86.36 per 10,000
person years at risk, corresponding to a more than 20-fold increased
risk of schizophrenia and a PAR of 0.12%. The broader category of a
parental history with mental illness was associated with risk of schi-
zophrenia (IRR = 2.07; 95% CI = 2.00–2.16) (PAR = 10.19%). We
found a PAR of similar magnitude (10.19%) associated with a parental
history of mental illness. Moreover, sex-specific cumulative incidences
(shown in Table 2) generally follow the same patterns as the IRR's.
Specifically, the lifetime risk of schizophrenia in male offspring where
both parents were diagnosed with schizophrenia was 20.33% (there
was not enough data to calculate the corresponding cumulative inci-
dence in women).
3.4. Parental age
We found that young maternal age was associated with offspring
risk (IRR = 1.19; 95% CI = 1.12–1.27; PAR = 1.56%) relative to the
reference category of age 25–29. No significant association was found
between increasing maternal age and increased risk of schizophrenia.
The risk of schizophrenia in the offspring increased with paternal age
(PA) at time of the child's birth. Relative to the age band 25–29 years
the IRR was 1.38; (95% CI = 1.25–1.51) and PAR 0.89% for PA
≥ 45 years. For PA 12–19 years we also found and increased relative
risk (IRR = 1.33, 95% CI 1.20–1.47). The sex-specific cumulative
incidences shown in Table 2 generally follow the same patterns as the
IRR. In male offspring the cumulative incidences were higher in both
ends of the paternal age distribution (PA categories ≥ 45 years and
12–19 years) while the effect of increasing paternal age had a more
linear pattern in female.
3.5. Parental loss
Table 1 shows that both early paternal loss (IRR = 1.16; 95%
CI = 1.11–1.21) (PAR = 2.34%) and early maternal loss (IRR = 1.20;
 H.J. Sørensen et al. / Schizophrenia Research 153 (2014) 214–219 217

Table 2 1.37; 95% CI = 1.33–1.41) than in women (CI = 0.91; 95% CI =0.88–
Cumulative incidencesa of schizophrenia at the 53rd birthday by gender, among 2,486,646 0.95). Birth in the capital was associated with a lifetime risk of 2.16
people born in Denmark 1955–1993, of which 17,389 developed schizophrenia
1970–2009.
(95% CI = 2.16–2.27) in men and 1.29 (95% CI = 1.19–1.39) in
women. In general the difference in cumulative incidences of schizo-
Males Females phrenia for men and women was found for all of the risk factors
Overall 1.37 (1.33–1.41) 0.91 (0.88–0.95) investigated.
Family history of mental illnessb
c
Both parents schizophrenic 20.23 (18.10, 22.62)
4. Discussion
One parent schizophrenic 9.53 (7.71, 11.79) 4.89 (4.50, 5.31)
c c
Both parents schizophrenia-like psychosis
One parent schizophrenia-like Psychosis 6.22 (4.91, 7.88) 3.41 (3.13, 3.71) To our knowledge, this is the first nation-wide population-based
Both parents psychiatric contact 6.29 (4.96, 7.96) 2.83 (2.59, 3.08) study which has simultaneously assessed the impact of several familial
One parent psychiatric contact 3.20 (2.47, 4.13) 1.80 (1.65, 1.97) and environmental risk factors for schizophrenia using 3 different
Parents with no psychiatric contact 1.19 (1.53, 1.23) 0.80 (0.77, 0.84)
epidemiologic measures. We were not able to identify an environmen-
Urbanicity
Capital 2.16 (2.06–2.27) 1.29 (1.19–1.39) tal (non-genetic) risk factor that can compete in magnitude with genet-
Capital suburb 1.62 (1.49–1.77) 1.09 (0.97–1.22) ic factors, as far as individual predictive power is concerned. Another
Provincial city 1.33 (1.25–1.42) 0.83 (0.76–0.91) main finding was that we were not able to suggest genetic factors that
Provincial town 1.17 (1.12–1.23) 0.80 (0.75–0.86)
could be targeted in any large-scale preventive effort. Additional find-
Rural area 1.01 (0.95–1.08) 0.76 (0.71–0.82)
Maternal loss ings were that all of the studied familial and environmental risk factors
Death 2.44 (2.09–2.85) 1.05 (0.84–1.31) were associated with schizophrenia and the highest population-
Alive 1.36 (1.32–1.40) 0.91 (0.88–0.94) attributable risk (PAR) was associated with urban birth (11.73%);
Paternal lossb followed by parental psychiatric contact PAR = 10.19%, whereas the
Death 2.53 (1.89–3.38) 1.35 (1.11–1.64)
lowest PAR was associated with having both parents diagnosed with
Alive 1.32 (1.28–1.36) 0.89 (0.85–0.92)
Immigrationb schizophrenia (0.12%). Intermediate PAR's were found for paternal
2nd generation immigrants, both 2.60 (2.20–3.08) 1.46 (1.05–2.02) loss (2.34%), maternal loss (1.56%), second-immigration status
2nd generation, mother 2.08 (1.86–2.32) 1.27 (1.09–1.47) (0.70%), advanced paternal age (0.89%), maternal age 12–19 years
2nd generation, father 2.43 (2.13–2.79) 1.45 (1.17–1.79)
(1.56%) and having one parent with schizophrenia (2.67%).
Native Danes 1.29 (1.25–1.33) 0.87 (0.84–0.91)
Maternal age
12–19 1.66 (1.53–1.81) 0.95 (0.84–1.08) 4.1. Strengths and limitations
20–24 1.33 (1.26–1.40) 0.91 (0.85–0.98)
25–29 1.28 (1.22–1.35) 0.88 (0.82–0.94) The primary strength of this study is the prospective design and the
30–34 1.38 (1.30–1.48) 0.94 (0.86–1.02)
population-based nationwide registers, ensuring the very large sample
≥35 1.54 (1.44–1.65) 0.95 (0.86–1.06)
Paternal ageb containing all people born in Denmark in the period from 1955 to
12–19 1.63 (1.42–1.87) 0.80 (0.66–0.97) 1993 and the ability to follow this cohort and their parents until death
20–24 1.43 (1.30–1.58) 0.90 (0.83–0.97) or emigration. Our analytic approach precluded bias due to emigration
25–29 1.23 (1.18–1.28) 0.85 (0.78–0.93)
and death from other causes, and all exposures were recorded indepen-
30–34 1.31 (1.23–1.38) 0.90 (0.84–0.97)
35–44 1.43 (1.35–1.52) 0.95 (0.88–1.03)
dently of the outcome and therefore were not subject to selection or
≥45 1.66 (1.49–1.85) 1.13 (0.97–1.32) recall bias. The extensive Danish registers allowed us to investigate
a joint influences on the risk for schizophrenia of both familial predispo-
The cumulative incidence measures the percentage of people who had developed
schizophrenia at a given age, taken into account that people may die or migrate. sitions to mental illness as well as factors often hypothesized to repre-
b
People with missing father were treated as a separate category in the analysis and sent environmental risk factors. Our risk model could have been
were not shown in the table. further elaborated by considering the possibility of interactions be-
c
Too few cases to obtain a reliable estimate.
tween risk factors. However, this was considered outside the scope of
the present study at this point. Limitations of the study include that
we were only able to investigate the risk factor information contained
in the Danish registers. Hence, unregistered parental history with men-
tal illness for instance could not be included in the study, which might
underestimate the effect of a family history with psychiatric disorders.
95% CI = 1.14–1.27) (PAR = 1.56%) was associated with offspring risk
of schizophrenia. The effects were of similar order of magnitude in both 4.2. Further discussion
genders (Table 2).
The approach with mutual adjustment for familial risk, as well as for
a range of environmental risk factors, may demonstrate to what extent
3.6. Immigration status
risk estimates are strengthened or attenuated in the context of other
factors. We found that the effect of family history of schizophrenia or
Compared to offspring where both parents were born in Denmark,
family history of mental disorder on the risk of schizophrenia was atten-
offspring whose both parents were immigrants was associated with ele-
uated to a certain minor degree in the context of other factors such as
vated relative risk of schizophrenia (IRR = 1.59; 95% CI = 1.42–1.77)
paternal and maternal age, paternal and maternal loss and urban birth
(PAR = 0.70%) with effects that were similar in both genders (Table 2).
second generation immigration status. On the other hand, our findings
demonstrate the robustness of family history and the other studied
3.7. Cumulative incidence of schizophrenia risk factors in predictive models of schizophrenia. In a wider perspective
one might need to consider the potential clinical usefulness of the esti-
Table 2 shows for the first time the sex-specific lifetime risks, mea- mates provided in our sample. A UK-based study has performed a study
sured as the cumulative incidence of schizophrenia at the 53th birthday, which is a theoretical exercise on preventable risk factors (Kirkbride
of schizophrenia for parental history of mental illness, urban place of et al., 2010); it estimated that up to 22% of all psychoses in England
birth, advanced paternal age, parental loss and immigration status. could be prevented if exposures associated with increased incidence
The overall life-time risk of schizophrenia was higher in men (CI = in ethnic minority populations could be removed. In contrast, our
218 H.J. Sørensen et al. / Schizophrenia Research 153 (2014) 214–219
findings illustrate that the predictive power on the individual level is far
from any level that would be clinically useful. Thus, the variables with
the highest predictive power (both parents diagnosed with schizophre-
nia) have very limited impact on the total population morbidity.
Some of the observed effects of the risk factors were expected and
replicated other studies and meta-analyses (El-Saadi et al., 2004; Miller
et al., 2011; Petersen et al., 2011). It is noteworthy, however, that a
small, but significant, effect of young maternal age on the risk of schizo-
phrenia in the offspring was observed in our sample even after adjust-
ment for a wide range of other risk factors. The significant impact of
paternal or maternal loss on the risk of schizophrenia in the offspring,
even after adjustment for psychiatric family history, extends and cor-
roborates previous studies on other psychiatric outcomes (Laursen
et al., 2007) but it is worth noting that the effect seems robust for
schizophrenia as well.
4.3. Future directions
As mentioned previously, it is uncertain whether genetic factors
can be usefully targeted in any large-scale preventive effort. In this
context, it is important to emphasize that the main determinant of
the cumulative incidence, and hence the power to predict individual
risk, is the magnitude of the IRR. The IRR's that we found to be asso-
ciated with familial and environmental risk factors, and hence the
power of a wide range of environmental risk factors to predict risk
on an individual level will probably not be able to equate or exceed
the predictive power associated with a parental history of schizo-
phrenia. Therefore, the power to predict risk on an individual level
would probably rarely be sufficient to mandate individual screening
and intervention. Moreover, our results clearly demonstrate that the
factors that potentially could reduce a substantial load of the disease
burden as measured by the PAR only have a negligible individual
predictive power. Obviously our study did by no means cover all po-
tentially relevant risk factors. Hence, one could also hope that for ex-
ample genetic data or other biomarkers would enhance the ability to
predict disease. Presently, a general intervention, not a high risk
strategy, would only be relevant to the extent that causal factors
eventually, and hopefully, are being identified and confirmed in fu-
ture studies. The incorporation of the study of important interactions
between genetic and environmental risk factors could also be con-
sidered in the future (McGrath et al., 2013). Utilizing the current
multiple risk factor approach could potentially include new identi-
fied risk factors for schizophrenia and this approach could facilitate
the development of high-risk interventions targeting people with
specific genetic vulnerability where additional exposure to an envi-
ronmental factor would possibly be deleterious in terms of added
schizophrenia risk. Future studies could consider inclusion of risk
factors with a potential to be modified. These factors could include
urban birth or migration history (Kirkbride et al., 2010) and a
range of possibly associated other environmental factors that
might be tangible markers for prevention. Such factors might be ex-
posure to prenatal infection (Mortensen et al., 2007; Sorensen et al.,
2009,) nutritional variables (Insel et al., 2008; McGrath et al., 2010;
Sorensen et al., 2011), or other potentially modifiable prenatal vari-
ables that have been linked with increased risk of schizophrenia
(Sorensen et al., 2004) or psychotic symptoms during adolescence
(Gunawardana et al., 2011). The incorporation of a wider range of
environmental risk factors might in the future improve the power
to predict individual risk of schizophrenia, for instance, by individual
scoring of exposure to environmental risk factors. In order to im-
prove prediction of schizophrenia, it may be necessary to incorpo-
rate the notion that there may be several pathways to the onset of
schizophrenia, putatively determined by a complex interplay of mul-
tiple factors across the life course (Van Os et al., 2008). At this point,
it may be concluded that in a nation-wide population-based study,
the simultaneously assessment of several familial and individual
risk factors for schizophrenia suggest limited predictive power on
the individual level. Moreover, the variables with the highest predic-
tive power have very limited impact on the total population morbid-
ity. Among the prerequisites for improved results in this field is the
continued search for causal factors that are prevalent in the general
population and factors where intervention is acceptable.
Contributors
Drs. Mortensen, Sørensen, Nielsen and Pedersen contributed to the design and dr.
Sørensen contributed to the initial preparation of the manuscript. Drs. Nordentoft and
Benros contributed by reviewing the manuscript and interpreting the data. Drs. Nielsen
and Pedersen performed the analyses and had full access to all of the data in the study
and take responsibility for the integrity of the data and the accuracy of the data analysis.
All authors approved the final version of the manuscript.
Conflict of interest
Neither of the authors declare any conflicts of interest.
Acknowledgment
The study was supported financially by:grants from The Lundbeck Foundation and
The Stanley Foundation. Neither funders had a role in the design and conduct of the
study, collection, management, analyses, and interpretation of the data, and preparation,
review, or approval of the manuscript and the decision to submit the manuscript for
publication.
References
Bruzzi, P., Green, S.B., Byar, D.P., Brinton, L.A., Schairer, C., 1985. Estimating the population
attributable risk for multiple risk factors using case–control data. Am. J. Epidemiol.
122, 904–914.
Cantor-Graae, E., Pedersen, C.B., 2007. Risk of schizophrenia in second-generation
immigrants: a Danish population-based cohort study. Psychol. Med. 37,
485–494.
Cantor-Graae, E., Pedersen, C.B., 2013. Full spectrum of psychiatric disorders related to
foreign migration: a Danish population-based cohort study. JAMA Psychiatry 70,
427–435.
Dean, K., Stevens, H., Mortensen, P.B., Murray, R.M., Walsh, E., Pedersen, C.B., 2010. Full
spectrum of psychiatric outcomes among offspring with parental history of mental
disorder. Arch. Gen. Psychiatry 67, 822–829.
El-Saadi, O., Pedersen, C.B., McNeil, T.F., Saha, S., Welham, J., O'Callaghan, E., Cantor-Graae,
E., Chant, D., Mortensen, P.B., McGrath, J., 2004. Paternal and maternal age as risk fac-
tors for psychosis: findings from Denmark, Sweden and Australia. Schizophr. Res. 67,
227–236.
Gottesman, I.I., Laursen, T.M., Bertelsen, A., Mortensen, P.B., 2010. Severe mental dis-
orders in offspring with 2 psychiatrically ill parents. Arch. Gen. Psychiatry 67,
252–257.
Granville-Grossman, K.L., 1966. Early bereavement and schizophrenia. Br. J. Psychiatry
112, 1027–1034.
Gunawardana, L., Zammit, S., Lewis, G., Gunnell, D., Hollis, C., Wolke, D., Harrison, G.,
2011. Examining the association between maternal analgesic use during preg-
nancy and risk of psychotic symptoms during adolescence. Schizophr. Res. 126,
220–225.
Insel, B.J., Schaefer, C.A., McKeague, I.W., Susser, E.S., Brown, A.S., 2008. Maternal iron de-
ficiency and the risk of schizophrenia in offspring. Arch. Gen. Psychiatry 65,
1136–1144.
Kendler, K.S., Diehl, S.R., 1993. The genetics of schizophrenia: a current, genetic-
epidemiologic perspective. Schizophr. Bull. 19, 261–285.
Kirkbride, J.B., Jones, P.B., 2011. The prevention of schizophrenia—what can we learn from
eco-epidemiology? Schizophr. Bull. 37, 262–271.
Kirkbride, J., Coid, J.W., Morgan, C., Fearon, P., Dazzan, P., Yang, M., Lloyd, T., Harrison,
G.L., Murray, R.M., Jones, P.B., 2010. Translating the epidemiology of psychosis
into public mental health: evidence, challenges and future prospects. J. Publ.
Ment. Health 9, 4–14.
Laursen, T.M., Munk-Olsen, T., Nordentoft, M., Bo, M.P., 2007. A comparison of selected
risk factors for unipolar depressive disorder, bipolar affective disorder, schizoaffective
disorder, and schizophrenia from a Danish population-based cohort. J. Clin. Psychia-
try 68, 1673–1681.
Matheson, S.L., Shepherd, A.M., Laurens, K.R., Carr, V.J., 2011. A systematic meta-review
grading the evidence for non-genetic risk factors and putative antecedents of schizo-
phrenia. Schizophr. Res. 133, 133–142.
McGrath, J.J., Eyles, D.W., Pedersen, C.B., Anderson, C., Ko, P., Burne, T.H., Norgaard-
Pedersen, B., Hougaard, D.M., Mortensen, P.B., 2010. Neonatal vitamin D status and
risk of schizophrenia: a population-based case–control study. Arch. Gen. Psychiatry
67, 889–894.
McGrath, J.J., Mortensen, P.B., Visscher, P.M., Wray, N.R., 2013. Where GWAS and epidemi-
ology meet: opportunities for the simultaneous study of genetic and environmental
risk factors in schizophrenia. Schizophr. Bull. 39, 955–959.
Miller, B., Messias, E., Miettunen, J., Alaraisanen, A., Jarvelin, M.R., Koponen, H., Rasanen, P.,
Isohanni, M., Kirkpatrick, B., 2011. Meta-analysis of paternal age and schizophrenia
risk in male versus female offspring. Schizophr. Bull. 37, 1039–1047.
 H.J. Sørensen et al. / Schizophrenia Research 153 (2014) 214–219
Mortensen, P.B., Norgaard-Pedersen, B., Waltoft, B.L., Sorensen, T.L., Hougaard, D., Yolken,
R.H., 2007. Early infections of Toxoplasma gondii and the later development of schizo-
phrenia. Schizophr. Bull. 33, 741–744.
Mortensen, P.B., Pedersen, M.G., Pedersen, C.B., 2010. Psychiatric family history and
schizophrenia risk in Denmark: which mental disorders are relevant? Psychol. Med.
40, 201–210.
Munk-Jorgensen, P., Mortensen, P.B., 1997. The Danish Psychiatric Central Register. Dan.
Med. Bull. 44, 82–84.
Pedersen, C.B., 2006. No evidence of time trends in the urban–rural differences in schizo-
phrenia risk among five million people born in Denmark from 1910 to 1986. Psychol.
Med. 36, 211–219.
Pedersen, C.B., Mortensen, P.B., 2001. Evidence of a dose–response relationship between
urbanicity during upbringing and schizophrenia risk. Arch. Gen. Psychiatry 58,
1039–1046.
Pedersen, C.B., Mortensen, P.B., 2006. Are the cause(s) responsible for urban–rural differ-
ences in schizophrenia risk rooted in families or in individuals? Am. J. Epidemiol. 163,
971–978.
Pedersen, C.B., Gotzsche, H., Moller, J.O., Mortensen, P.B., 2006. The Danish Civil
Registration System. A cohort of eight million persons. Dan. Med. Bull. 53,
441–449.
219
Petersen, L., Mortensen, P.B., Pedersen, C.B., 2011. Paternal age at birth of first child and
risk of schizophrenia. Am. J. Psychiatry 168, 82–88.
Sorensen, H.J., Mortensen, E.L., Reinisch, J.M., Mednick, S.A., 2004. Association between pre-
natal exposure to analgesics and risk of schizophrenia. Br. J. Psychiatry 185, 366–371.
Sorensen, H.J., Mortensen, E.L., Reinisch, J.M., Mednick, S.A., 2009. Association between
prenatal exposure to bacterial infection and risk of schizophrenia. Schizophr. Bull.
35, 631–637.
Sorensen, H.J., Nielsen, P.R., Pedersen, C.B., Mortensen, P.B., 2011. Association between
prepartum maternal iron deficiency and offspring risk of schizophrenia:
population-based cohort study with linkage of Danish national registers. Schizophr.
Bull. 37, 982–987.
Van Os, J., Rutten, B.P., Poulton, R., 2008. Gene-environment interactions in schizophrenia:
review of epidemiological findings and future directions. Schizophr. Bull. 34, 1066–1082.
Weiser, M., Werbeloff, N., Vishna, T., Yoffe, R., Lubin, G., Shmushkevitch, M., Davidson, M.,
2008. Elaboration on immigration and risk for schizophrenia. Psychol. Med. 38,
1113–1119.
World Health Organization, 1967. Manual of the International Statistical Classification of
Diseases, Injuries, and Causes of Death (ICD-8). WHO, Geneva.
World Health Organization, 1992. International Statistical Classification of Diseases and
Related Health Problems (ICD-10). WHO, Geneva.