Birth cohort designs are useful in studying adult disease trajectories and outcomes, such as schizophrenia. We review the schizophrenia research performed in the Northern Finland Birth Cohort 1966 (NFBC 1966), which includes 10,934 individuals living in Finland at 16 years of age who have been monitored since each mother’s mid-pregnancy. By the age of 44, 150 (1.4%) had developed schizophrenia. There are 77 original papers on schizophrenia published from the NFBC 1966. The early studies have found various risk factors for schizophrenia, especially related to pregnancy and perinatal phase. Psychiatric and somatic outcomes were heterogeneous, but relatively poor. Mortality in schizophrenia is high, especially due to suicides. Several early predictors of outcomes have also been found. Individuals with schizophrenia have alterations in brain morphometry and neurocognition, and our latest studies have found that the use of high lifetime doses of antipsychotics associated with these changes. The schizophrenia research in the NFBC 1966 has been especially active for 20 years, the prospective study design and long follow-up enabling several clinically and epidemiologically important findings. When compared to other birth cohorts, the research in the NFBC 1966 has offered also unique findings on course and outcome of schizophrenia.
Birth cohort studies were originally designed to study early health outcomes, but as the earliest birth cohorts started in the 1940s to the 1960s have matured, adult outcomes have also been studied [
The Northern Finland Birth Cohort 1966 (NFBC 1966) is a birth cohort which has been studied extensively with almost 1000 peer-review publications in various fields of medicine (
In this paper, we systematically review the schizophrenia research performed in the NFBC 1966.
The literature search was completed in PubMed in September 2014. The keywords included “Northern Finland Birth Cohort” and words related to schizophrenia (including schizotypy and psychotic disorders). Additionally, we manually searched for articles from the extensive publication list maintained by the NFBC 1966 (
The sample was drawn from the NFBC 1966, which is based on 12,068 pregnant women and their 12,058 live-born babies in two northernmost Finnish provinces (Oulu and Lapland), with an expected date of birth during 1966 [
The 31-year follow-up of the whole cohort was conducted in 1997. Comprehensive data on somatic and mental health and the use of health services and medications as well as socioeconomic status were collected using a postal questionnaire (
The 34-year follow-up for psychoses was conducted in 1999–2001. Cohort members with a psychotic episode by the end of the year 1997 (cases) were detected using the Finnish Hospital Discharge Register (FHDR), currently known as Care Register for Health Care. The control subjects were selected randomly from the cohort members without a psychotic episode, living in the region of Oulu. The 34-year follow-up consisted of structural magnetic resonance imaging (MRI) of the brain (GE Signa 1.5T scanner), measures of cognitive functioning, psychiatric interviews, and questions related to, for example, the use of antipsychotic medication, social background, and substance use. In the 34-year follow-up, 91 cases with a psychotic disorder (61 with schizophrenia) and 104 controls participated and provided written informed consent [
The 43-year follow-up for psychoses was conducted in 2008–2010. In addition to those with regard to whom a psychosis had been detected during the 34-year follow-up, those who had developed a psychosis between the years 1998 and 2008 according to the nationwide registers or self-report were invited to participate. The new control subjects were selected randomly from all the cohort members. The procedure was extended from the 34-year follow-up with functional MRI and more measures of cognitive functioning. Out of the eligible subjects, 107 cases with a psychotic disorder (54 with schizophrenia) and 194 controls participated and provided written informed consent. Of the participants, 54 cases with a psychotic disorder (40 with schizophrenia) and 76 controls participated, both at ages 34 and 43 [
The diagnoses of the participants of the 34-year follow-up were validated in accordance with DSM-III-R [
The first schizophrenia studies in the NFBC 1966 focused on risk factors. The main results of the studies are presented in Table
Risk factors of schizophrenia in the Northern Finland Birth Cohort 1966.
Risk factors | Risk estimates |
---|---|
Pregnancy and birth | |
Male gender | Risk of schizophrenia OR 1.8 (95% CI 1.1–3.0) [ |
Parental psychosis | Risk of schizophrenia HR 2.8 (95% CI 1.7–4.5) [ |
Maternal antenatal depression | Nonsignificant risk of schizophrenia RR 1.5 (95% CI 0.9–2.4) [ |
Unwantedness of a pregnancy | Risk of schizophrenia OR 2.4 (95% CI 1.2–4.8) [ |
High paternal social class at birth (females) | Risk of schizophrenia RR 2.4 (95% CI 1.2–4.9) [ |
Perinatal brain damage | Risk of schizophrenia OR 4.6 (95% CI 1.7–12.1) [ |
Deviant intrauterine growth | Increased risk of schizophrenia for both low (OR 2.5; 95% CI 1.2–5.1) and high (OR 2.4; 95% CI 1.1–4.9) birth weight and for both low (OR 2.6; 95% CI 1.1–5.9) and high (OR 1.8; 95% CI 1.0–3.5) birth length [ |
Birth order | Risk of schizophrenia among male first-borns (observed-expected ratio 1.5; 95% CI 1.0–2.2) and nonsignificant risk among female last-borns (1.3; 95% CI 0.9–1.9) [ |
Grand multiparity, six or more | Nonsignificant risk of schizophrenia OR 1.1 (95% CI 0.6–2.1), significant risk of other psychosis OR 2.3 (95% CI 1.2–4.7) [ |
|
|
Childhood and adolescence | |
Use of vitamin D supplementation, at least 2000 IU | Decreased risk of schizophrenia RR range: 0.08–0.23 [ |
Markers of development in childhood | Risk of schizophrenia, later age of learning to stand (RR 1.5; 95% CI 1.0–2.2) and walk (RR 1.3; 95% CI 1.0–1.6) [ |
Family type (single-parent versus two-parent) at birth to 14 years | No significant differences [ |
Viral central nervous system infection before age 14 | Risk of schizophrenia OR 4.8 (95% CI 1.6–14.0) [ |
Intelligence Quotient (IQ) <85 up to the age of 14 years | Risk of schizophrenia OR 4.8 (95% CI 2.2–10.3) [ |
Excellent school performance age 16 years (males) | Risk of schizophrenia OR 3.8 (95% CI 1.6–9.3) [ |
Lower school marks and below normal class age 16 years | Lower school marks associated with risk for nonpsychotic disorders, but not for schizophrenia or other psychoses. Below normal class associated with schizophrenia risk in boys OR 2.8 (95% CI 1.5–5.7) [ |
|
|
Other premorbid factors | |
Severe injuries (e.g., fractures) before onset of psychosis | Risk of psychotic disorder, having a fracture HR 2.9 (95% CI 1.4–6.0) [ |
Seasonality | Rate of first admission for schizophrenia lower in spring compared to other seasons [ |
Paternal age | No significant risk by paternal age categories [ |
|
|
Interaction studies | |
Parental psychosis and maternal antenatal depression | Risk of schizophrenia highest in the offspring with both maternal depressed mood during pregnancy and parental psychosis (OR 9.4; 95% CI 4.2–20.9) [ |
Parental psychosis and early risk factors | Risk for schizophrenia highest among those with parental psychosis and biological risk factor: parental psychosis and high birth weight HR 7.9 (95% CI 1.8–34.8), parental psychosis and high birth length HR 4.3 (95% CI 1.1–16.2), parental psychosis and any biological risk HR 3.6 (95% CI 1.3–10.3), and parental psychosis and high maternal education HR 0.2 (95% CI 0.1–0.9) [ |
|
|
Associations between risk factors | |
Motor performance at age 1 and school performance at age 16 | Age of learning to stand associated with school mark of physical education in schizophrenia but not in controls [ |
Parental psychosis and advanced paternal age | Maternal schizophrenia associated with higher advanced paternal age [ |
OR = odds ratio, CI = confidence interval, HR = hazard ratio, RR = risk ratio, and IU = international unit.
The most essential risk factors identified in the latest NFBC 1966 review [
Later, other risk factors have also been found, such as severe injuries [
In each stage of human development, there are risk factors influencing the occurrence of schizophrenia. It is important to identify the risk and create awareness among health professionals and decision makers to develop and implement preventive strategies, though it should be noted that most of the effect sizes detected are small or medium level. In schizophrenia risk studies in general, the risk factors with highest quality evidence have shown medium effect sizes [
In the NFBC 1966 on average of ten years from the onset of illness, the outcomes were heterogeneous, but relatively poor [
At the age of 34, outcomes were heterogeneous and varied largely, depending on the measures of outcomes. More than half (56%) of the individuals with schizophrenia were on disability pension [
There are several relevant predictors of outcomes, but especially factors related to illness onset are significant to course of illness. Statistically significant predictors of poor outcomes are presented in Table
Predictors of poor outcomes in schizophrenia in the Northern Finland Birth Cohort 1966.
Definition of poor outcome | Predictor of poor outcome |
---|---|
Poor clinical outcomes | |
(i) More hospitalizations, rehospitalization after first episode [ |
Family history of psychosis |
(ii) More symptoms, lack of remission [ |
Adolescence factors: smoking at age 14 years, poorer school performance at high school |
(iii) Treatment resistance [ |
Higher birth length and weight |
|
|
Poor social outcomes | |
(i) Poor occupational functioning |
Childhood and adolescence factors: lack of friends at childhood, poorer school performance at high school |
DUP = duration of untreated psychosis, SOFAS = Social and Occupational Functioning Assessment Scale.
Clinical and functional course of illness associates also with biological markers: earlier motor development at age one [
Differences between subjects with schizophrenia and controls in brain volumes and densities in several brain regions were found at the 34-year follow-up. Subjects with schizophrenia showed 2-3% lower whole brain, grey matter (GM), and white matter (WM) volumes and 7% higher cerebrospinal fluid volume when compared to controls [
The volume of hippocampus was reduced in individuals with schizophrenia or any psychosis when compared to the controls, but the difference was explained by the total brain volume reduction. In the amygdala, significant volume differences between individuals with schizophrenia and controls were not found [
At age 34, associations between GM and WM densities and earlier onset age (i.e., longer duration of illness) were found [
Between 34 and 43 years of age, Veijola et al. [
Neurocognitive performance, longitudinal change in cognition, and their associations with infant motor developmental markers, brain volume change, and antipsychotic medication have all been studied in the NFBC 1966 in schizophrenia in comparison with nonpsychotic controls.
Delayed infant motor development was associated with poorer adult cognition in schizophrenia at the age of 34 years in executive functions, verbal learning, and visuospatial working memory, but not in visual learning [
Age of learning to stand in infancy significantly inversely predicted later deterioration of executive functions with memory in adult schizophrenia between 34 and 43 years of age [
Global cognitive functioning as well as performance in domains of executive functions, working memory, and visual and verbal memory were lower in subjects with schizophrenia in comparison to control subjects cross-sectionally, both at the age of 34 [
Higher lifetime antipsychotic medication exposure was associated with poorer verbal learning and memory performance at the age of 34 and greater decline in verbal learning and memory between the ages of 34 and 43 years [
In the NFBC 1966, the proportion of smokers was about 50% among people with schizophrenia and the initiation of regular smoking was closely associated with the onset of the disorder [
In addition, in the NFBC 1966, schizophrenia has been associated strongly with epilepsy by the age of 28 years, neurological diseases in general, and inflammatory bowel diseases [
In the NFBC 1966, the overall mortality rate among subjects with schizophrenia was 14% before the age of 27 years, with half of the deaths being suicides [
Those with psychotic disorders exhibited higher risks of criminal behavior [
In the NFBC 1966, the associations between the use of antipsychotic medication and related factors have been extensively studied [
Factors associated with the use of antipsychotic medication in the Northern Finland Birth Cohort 1966.
Factors | Effect of higher use of antipsychotic medication |
---|---|
Sex | Females used more often medication [ |
Age at onset of psychosis | No association [ |
Psychiatric hospital treatment | Increased need of hospital treatment [ |
Clinical measures (PANSS, CGI, SOFAS, and remission) | Poorer clinical and social outcome [ |
Occupational functioning | Higher rate of disability pensions and lower rate of employment [ |
Brain volume loss | Increased brain loss [ |
Change of cognition | Decline of verbal learning and memory [ |
Suicidal ideation | No association when symptoms of depression and anxiety were controlled for [ |
Hyperlipidemia | Increased risk of hyperlipidemia [ |
PANSS = Positive and Negative Syndrome Scale, CGI = Clinical Global Impression, and SOFAS = Social and Occupational Functioning Assessment Scale.
The use of antipsychotic medication predicted an overall poorer outcome; for example, subjects using antipsychotics needed more hospital treatment, had poorer clinical and social outcomes, and were more often unemployed or on disability pension [
Higher cumulative antipsychotic medication exposure (chlorpromazine equivalent dose-years) over the 9-year follow-up between the two brain scans was associated with brain volume reduction [
Schizophrenia is a highly heritable disease. Recently, the Schizophrenia Working Group of the Psychiatric Genomics Consortium [
The cohort studies have focused on psychometric properties on these scales and differences in the traits by gender and education [
The genetic studies in the NFBC 1966 have focused especially on DISC1 (disrupted-in-schizophrenia 1), finding associations between DISC1 and social anhedonia [
The NFBC 1966 has been especially active for 20 years. The prospective study design and long follow-up have enabled several clinically and epidemiologically important findings. The early studies in the NFBC 1966 focused on risk factors for schizophrenia. Various risk factors have been found, related to, for example, family history of psychosis, adverse events during pregnancy and delivery, and delays in early developmental milestones. Psychiatric course of illness and outcomes are heterogeneous and unfortunately relatively poor. There are several predictors for outcomes, one of the most important being early age of illness onset. Some of the predictors of outcomes may be modified, such as substance use and duration of untreated psychosis.
There are differences between subjects with schizophrenia and controls in brain volume and densities in several brain areas. Global cognitive functioning and performance in various cognitive domains were worse in schizophrenia than in the controls cross-sectionally, but longitudinally cognition mostly followed the normal course of aging in midlife. High doses of antipsychotics associated with both brain volume loss and decline in cognition, and also more severe illness in terms of symptomatology, the need of hospitalization, and functional outcomes. Additionally, the use of antipsychotics was related to hyperlipidemia. Physical diseases, such as metabolic syndrome and epilepsy, are common in schizophrenia, and mortality is high, especially due to suicides.
The NFBC 1966 has provided important knowledge on epidemiology and etiology of schizophrenia. In particular, the possibility of prospectively and reliably analyzing the very early risk factors of schizophrenia is unique. Additionally, the results from the last ten years have brought information that can be utilized in developing clinical practice as well. For example, results on psychiatric and somatic outcomes, mortality, and predictors of outcomes are important with regard to the health promotion and clinical care in schizophrenia. The results also indicate that suicide prevention, early detection and intervention, monitoring somatic health, and active treatment of metabolic abnormalities are highly important. The sample is naturalistic, describing the real-life situation, so the results are generalizable to other settings. The NFBC 1966 studies have analysed also methodological questions, such as diagnostic validity [
The especially interesting and novel findings of the NFBC 1966 are those concerning potentially harmful effects of high doses of antipsychotics on changes in brain morphometry and cognitive functioning [
The NFBC 1966 is not the only birth cohort that has studied schizophrenia. Welham et al. [
Studies on cognition on other birth cohorts have focused mostly on premorbid cognition. MacCabe [
To our knowledge only one other birth cohort study has focused on brain MRI, showing structural alterations of the cerebral cortex, particularly in the frontal and temporal lobes, and ventricular enlargement in schizophrenia [
Somatic comorbidities have been studied in the Helsinki Birth Cohort Study, where individuals with schizophrenia had a higher risk of hospitalization and mortality for coronary heart disease; however, patients with schizophrenia used less lipid-lowering and antihypertensive drug treatment [
The register-based Danish national birth cohort has been utilized when studying family history and variation in single-nucleotide polymorphisms to risk of schizophrenia. Agerbo et al. [
There is also younger birth cohort based on the same geographical area than the NFBC 1966; the studies relating to schizophrenia in the Northern Finland Birth Cohort 1986 have so far studied mainly those with clinical or familial risk for psychosis or schizophrenia [
In all, many of the risk factor findings from the NFBC 1966 have been replicated in other birth cohorts. To our knowledge, other birth cohort studies have not analyzed clinical and occupational course of illness and their determinants after onset of schizophrenia, though these kind of clinically relevant studies would be especially important in epidemiologically sound samples like birth cohorts. There is a lack of birth cohort studies with straight clinical implications, including postmorbid data, and this should be corrected in the future as birth cohorts get older.
The strengths of the birth cohort samples in general and also of the NFBC 1966 include the prospective design and large number of unselected matched controls. This design provides unique possibilities of analyzing causal relationships, for example, between risk factors during pregnancy and delivery and later risk of illness. The major benefit in the NFBC 1966 is also the ability to utilize large nationwide health registers [
The NFBC 1966 and other birth cohort studies have various limitations. Participants are not randomly assigned to exposure; thus, any observed associations may be confounded by unmeasured factors. For example, very sick individuals get more antipsychotics than persons with minimal symptomatology. Also, naturally, birth cohort studies include only case and outcome data until the point of data collection used in each study. In the birth cohort studies, the number of cases is often relatively small and power to detect even moderate signal may be low. Additionally, originally birth cohort was not planned for schizophrenia studies specifically: this means that all relevant data regarding schizophrenia research has not been collected (e.g., childhood behaviour and symptoms or treatment aspects, especially regarding psychosocial treatments).
Schizophrenia research has been conducted for over 20 years in the NFBC 1966. Several interesting and unique findings regarding risk factors, somatic and psychiatric outcomes, brain morphology and cognition, antipsychotics use, and genetics have been reported. In the future, with the cohort members aging, somatic health, consequences of the use of antipsychotics, and the long-term psychiatric and functional outcomes will be important to explore.
The funders had no role in the design or conduct of the study.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This study has been funded by the Academy of Finland (no. 132071, no. 268336, and no. 278286), Brain and Behavior Research Foundation, Jalmari and Rauha Ahokas Foundation, Sigrid Jusélius Foundation, and Northern Finland Health Care Support Foundation. The authors would like to acknowledge all the participants in this study and researchers collecting the data.