Associations between early development and outcome in schizophrenia — A 35-year follow-up of the Northern Finland 1966 Birth Cohort

Available online at www.sciencedirect.com

Schizophrenia Research 99 (2008) 29 – 37 www.elsevier.com/locate/schres

Associations between early development and outcome in schizophrenia — A 35-year follow-up of the Northern Finland 1966 Birth Cohort Erika Jääskeläinen a,⁎, Jouko Miettunen a,b , Juha Veijola a,b , John J. McGrath c , Graham K. Murray d , Peter B. Jones d , Matti Isohanni a,e a

e

Department of Psychiatry, University of Oulu, Finland, P.O. BOX 5000, 90014 University of Oulu, Finland b Academy of Finland, P.O. Box 99, 00501 Helsinki, Finland c Queensland Centre for Mental Health Research, Department of Psychiatry, University of Queensland, Wacol, 4076, Australia d Department of Psychiatry, University of Cambridge, Cambridge CB2QQ, UK Department of Public Health Science and General Practice, University of Oulu, P.O. BOX 5000, 90014 University of Oulu, Finland Received 15 February 2007; received in revised form 13 November 2007; accepted 18 November 2007 Available online 3 January 2008

Abstract Delayed neuromotor development carries an increased risk of developing schizophrenia, and some authors have assumed that risk factors for schizophrenia such as delayed development are also prognostic indicators for patients with established illness. In those who do develop schizophrenia, it is not clear if these same early developmental markers influence the outcome of illness. Our aim was to examine the association between infant developmental milestones and a range of outcomes in patients with schizophrenia. Our sample was drawn from Northern Finland 1966 Birth Cohort and included 109 subjects for whom prospectively collected information on age of learning to stand, walk and talk was available and who had developed schizophrenia by the age 35 years. By utilizing national registers we examined outcomes related to service utilization, educational achievement, and occupational status. Age of illness onset was also analyzed. Based on the diagnostic interview, a subgroup of 59 cases was assessed in clinical examinations on functioning and quality of life. Contrary to a widespread assumption within the field of schizophrenia research, later attainment of developmental milestones was not associated with poor outcome. We conclude that risk factors for schizophrenia are not necessarily prognostic factors. © 2007 Elsevier B.V. All rights reserved. Keywords: Schizophrenia; Development; Outcome; Course of illness; Predictor

1. Introduction There is convincing evidence that children who have delays in neurocognitive and motor development ⁎ Corresponding author. Tel.: +358 40 7474376; fax: +358 8 333167. E-mail address: [email protected] (E. Jääskeläinen). 0920-9964/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2007.11.024

have a small but significantly increased risk of developing schizophrenia (Fish et al., 1992; Jones et al., 1994; Isohanni et al., 2001; Cannon et al., 2002). However, once illness develops, the association between these developmental markers or risk factors and the subsequent course of illness remains poorly understood.

30

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

Some authors and clinicians have hypothesized, that risk factors for schizophrenia such as delayed development are also prognostic indicators for patients with established schizophrenia. For example Johnstone and colleagues (1995) stated that “the most malignant form of schizophrenia is neurodevelopmental”, while Castle et al. (1993) argued that “young males are especially prone to a severe neurodevelopmental form of illness associated with premorbid deficits”. The most specific hypothesis of this nature comes from Murray et al. (1992), who conjectured that “congenital schizophrenia is a consequence of aberrant brain development during fetal and neonatal life. Such patients show structural brain changes and cognitive impairment, and in their male predominance, early onset, and poor outcome, they reflect Kraepelin's original description of dementia praecox.” In spite of the above assertions, no studies analyzing associations between prospectively collected infant development and clinical outcomes in schizophrenia patients have been published. Some studies have used proxy measures of developmental deviance such as premorbid unemployment, single marital status and poor academic achievement to make inferences about neurodevelopment (e.g. van Os et al., 1995). Other researchers have utilized retrospectively collected information about development to examine associations with outcomes, and therefore may be vulnerable to recall bias. For example, Jonsson and Nyman (1984, 1991) examined past history as retrospectively recorded in clinical files to study 110 first admitted schizophrenia patients in Sweden. They found that the presence of “neurological symptoms” (e.g. bad coordination of muscles or speech, difficulties in writing and reading) during childhood and adolescence was associated with poorer social outcome. Rossi et al. (2000) studied 32 schizophrenia outpatients in Italy and explored developmental trajectories relying on maternal recall. They found that those patients who had initially high and stable premorbid behavioral abnormalities also had more severe negative symptoms than patients who had slight but increasing behavioral abnormalities. We are not aware of any prior studies examining the association between prospectively collected measures of early infant development and later clinical and social outcomes among patients with schizophrenia, and therefore we aimed to conduct such a study. We hypothesized that risk factors for schizophrenia also are risks for more severe disease; within individuals who develop schizophrenia, those cases with later developmental milestones (standing, walking, talking)

would have less favorable outcomes, compared with early learners. 2. Methods 2.1. Subjects and data collection The NFBC 1966 is based upon 12 058 live-born children in the provinces of Lapland and Oulu with an expected delivery date during 1966, representing 96% of all births (Rantakallio, 1969; Isohanni et al., 1997). Data on biological, socioeconomic and health conditions, living habits and family characteristics were collected prospectively from pregnancy up to the age of 31. There were 11 017 eligible individuals in Finland at the age of 16 years. The Faculty of Medicine Ethics Committee of the University of Oulu has approved and keeps the study design of the NFBC 1966 under review. In the field surveys conducted in 1997 and in 1999–2001, subjects were given a complete description of the study. All subjects had an opportunity to refuse to participate and gave written informed consent. A total of 83 individuals did not consent to the use of their data and have been excluded, as well as seven organic mental disorders, leaving 10 934 cohort members for the present study (5586 boys, 51%). 2.2. Predictor variables — developmental milestones Developmental data originally obtained during children's visits to welfare centers (the average cohort member made 10 such visits in the first 12 months of life) were collected and supplemented with a special examination performed at age approximately 1 year. In 96% of the cases the information was collected at an age of 11.5 months or later (Rantakallio et al., 1985). The following developmental milestones were examined at this one-year examination and utilized in this study: 1. Age of standing without support and walking with and without support. Parents were asked the age (in full months) that the child learned these skills. We used the following categories for analyses: standing without support (at 10 months or before vs. at 11 months or later) and walking without support (at 11 months or before vs. at 12 months or later). 2. Age of learning to talk. Parents were asked how many words the child spoke by the age 12 months. For analyses following categories were used: at least

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

three words vs. none or at most two words by the age 12 months. The cutoffs for developmental variables were based on distribution of these same variables in controls (i.e. cohort members without any psychiatric diagnosis) (Table 1). For each developmental variable we chose the cutoff that best divided the control group into half. This information is presented in more detail in Isohanni et al. (2001). In addition to these, in order to obtain a continuous (as opposed to ordinal or categorical) developmental independent variable, we used an Infant Motor Development (IMD) summary score (described in detail by Murray et al., 2006; Ridler et al., 2006). We applied Principal Component Analysis to the three measures of motor development for each study participants: age of learning to stand without support (months), learning to walk with support (months) and learning to walk without support (months). The first Principal Component was defined as a summary measure of infant motor development (IMD). Higher IMD scores indicate delayed motor milestones (Murray et al., 2006). The first Principal Component explained 72.2% of the total variance in the above described milestone variables. Some individuals were unable to perform standing and/or walking at the time of the developmental assessment so in order to calculate developmental measures, we imputed their standing or walking ages using an a priori conservative method: We attributed a value equal to the age at the time of the developmental assessment plus one month (Murray et al., 2006).

Table 1 The distribution of developmental variables in subjects with schizophrenia and controls Developmental variable Standing By 10 months By 11 months 12 months or later Walking By 10 months By 11 months 12 months or later Talking by 12 months No words 1–2 words 3 or more words

Controls

n%

Subjects with schizophrenia

n%

4074 2430 2969

43% 26% 31%

24 25 46

25% 26% 48%

1840 2286 5333

20% 24% 56%

8 19 68

8% 20% 72%

2169 2033 5281

23% 21% 56%

27 21 50

28% 21% 51%

31

2.3. Case identification The nationwide Finnish Hospital Discharge Register (FHDR) covers all mental and general hospitals, as well as beds in local health centers and private hospitals nationwide. All cohort members over 16 years appearing on the FHDR until the end of 1997 for any mental disorder (i.e. ICD-8 290–309, ICD-9 diagnoses 290– 316, and ICD-10 F00–F69, F99) were identified. All case records were scrutinized and diagnoses were validated for the DSM-III-R criteria. The reliability of researchers assessing schizophrenia diagnoses was good (kappa = 0.85). Altogether 160 subjects with known psychotic episodes until the year 1999 were detected. The detection of subjects and validation of diagnoses are described in Isohanni et al. (1997) and Moilanen et al. (2003). Of these 160 subjects, 14 (8.8%) had died by the year 2001, ten of these had schizophrenia (Alaräisänen et al., 2006). During 1999–2001 all 146 living subjects (84 males, 58%) with at least one known psychotic episode in their life were invited to participate in a field study. Altogether 91 (62%) subjects with history of any psychosis participated and all participants gave written informed consent. There were 40 living subjects with schizophrenia who did not participate in the field study. The Structured Clinical Interview for DSM-III-R, the SCID (Spitzer et al., 1989) was used for diagnostic assessment for the participating subjects with history of psychosis, together with all available anamnestic information including individual hospital notes. After diagnostic interviews a total of 59 cases with schizophrenia (34 men, 58%) were detected and included in this study. Schizophrenia included following DSM-III-R diagnoses: 295.1; 295.3; 295.6; and 295.9. The field study in 1999–2001 is presented in detail in Tanskanen et al. (2005) and Ridler et al. (2006). The analyses for early development and outcome in this study were made in two categories: 1) Register-based information about outcome was used for all individuals at any time fulfilling the DSM-IIIR criteria for schizophrenia (N = 109). This included those schizophrenia subjects from the field study in 1999–2001 (N = 59), but also those detected by the year 1999 but who did not participate in the field study in 1999–2001 (N = 40). Deceased schizophrenia cases (N = 10) were also included. 2) In addition, data from assessments based on personal interviews were used for a subset of schizophrenia cases diagnosed and interviewed at the field study in 1999–2001 (N = 59).

32

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

2.4. Measures of outcome 2.4.1. Functioning and quality of life by the age of 35 years Of those 59 cases who agreed to participate in the field study and interviews in 1999–2001, we obtained the following measures: 1. SOFAS (Social and Occupational Functioning Assessment Scale) score, derived from the Global Assessment Scale (GAS), was used to rate the social and occupational functioning of the patients (Spitzer et al., 2000) during the previous week. Higher scores reflect better functioning. 2. 15D is a generic, standardized, self-administered measure of health-related quality of life. It has 15 separated dimensions, assessing physical, mental and social well-being (Sintonen, 2001). The dimensions include: breathing, speech, vision, mobility, usual activities, vitality, hearing, eating, elimination, sleeping, distress, discomfort and symptoms, sexual activity, depression and mental function. A single index number is generated from the scores and this number ranges from 0 to 1, with high scores meaning good quality of life. 2.4.2. Inpatient psychiatric hospitalizations Though number of hospitalizations is an imperfect measure of severity of illness, it is widely used as a proxy measure in several earlier outcome studies (e.g. Suvisaari et al., 1998; Harrison et al., 2001). In order to explore the association between developmental milestones and clinical outcomes we examined a measure of service utilization based on national register databases for all the 109 cases (all hospital data were collected till the year 2000 from the Finnish Hospital Discharge Register): 1. Cumulative number of days in psychiatric hospital. 2. Proportion of time spent in hospital after first psychiatric diagnosis (%) per patient. 2.4.3. Other measures of general outcomes 1. Age at illness onset for psychosis. Ascertained from medical records and defined as the age when first experiencing psychotic symptoms (Räsänen et al., 1999). This variable was available for all the 109 cases. 2. Working ability. Since the year 1996 on disability pension vs. not. Data from the Social Insurance Institution of Finland. This variable was available for all the 109 cases.

3. Marital status. In the field study in 1999–2001 cases were asked about their marital status. The variable was in two categories: married or cohabiting vs. divorced, not married or not cohabiting. This variable was available for 59 cases who participated in the field study in 1999–2001. 4. Educational level. The educational level that the subject had attained by the year 1997: information from the National Educational Registry of Statistics Finland. The variable was in two categories: secondary or tertiary education (over 9 years) vs. basic education (9 years or less). This variable was available for all the 109 cases. 2.5. Potential confounding and mediating factors 1. Sex. Men vs. women. 2. Perinatal risk. Yes vs. no. Obstetric complications include low birth weight (b2500 g) or short gestation (b37 weeks) or perinatal brain damage (Jones et al., 1998). 3. Family history of psychosis. In the case of the cohort members with schizophrenia, all hospital notes and available outpatient notes and parents' hospitalizations during 1972–2000 from the FHDR were checked to find out whether or not a first degree relative had experienced a psychotic episode. Additionally, in the field survey during 1999–2001 all of the participating patients underwent Family Interview for Genetic Studies (FIGS; Maxwell, 1992). In the FIGS procedure the patient and the mother, or if the mother was not willing or able to participate, the father or one of the siblings, were asked whether any first degree relatives had experienced psychotic symptoms. 4. Father's social class in 1980. Paternal social classes I–II (high) vs. III–V (low). Information was collected from the questionnaire for child, parents or school nurses at the child's age of 14 years. Classes I–II included those with the highest professions, III–IV included skilled and unskilled workers, respectively, and V included farmers. 5. Age at illness onset for psychosis. Age (years) when first psychotic symptoms emerged (Räsänen et al., 1999). Taking the period effect into account is important, as there has been a notable decrease from 1982 to 1992 in the number of hospital beds in Finland during the study period (Tuori et al., 1998). In order to decrease the period effect and the effect of length of illness, the age of illness onset was used as covariate (analyzed as continuous) when analyzing association between developmental variables and

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

number of hospital days and proportion of time spent in hospital. 2.6. Statistical analyses Statistical analyses were made with SPSS 15.0 for Windows. Outcome variables were dichotomized. A median as a cutoff point was used for continuous outcome variables. Differences in outcomes between ages learning to stand (divided in two strata: ‘early’ = 10 months or less vs. ‘late’ = at 11 months or later), walk (‘early’ = 11 months or less vs. ‘late’ = at 12 months or later), and talk (‘early’ = at least three words vs. ‘late’ = none or at most two words by the age 12 months) were analysed with cross-tabulations.

33

The relationships between the variables of interest (ages of learning to stand, walk and talk and IMD and outcome variables) were assessed with and without adjustment for potential confounding by sex, obstetric complications, family history of psychosis, father's social class, and age of illness onset by logistic regression model. 2.7. Missing data Data on standing, walking and talking were missing from the 1-year examination in 10% to 13% of the cases with schizophrenia. Missing data on outcome measures were at most 5% and on potential confounders 0%.

Table 2 The age of achieving developmental milestones and the course of illness in schizophrenia Outcome variables ⁎ Learned to stand (months) Under median age of illness onset (n %) Over median psychiatric treatment days (n %) Over median time spent (%) in hospital (n %) Basic educational level (n %) Pensioned (n %) Learned to walk (months) Under median age of illness onset (n %) Over median psychiatric treatment days (n %) Over median time spent (%) in hospital (n %) Basic educational level (n %) Pensioned (n %) Talking by 12 months

Under median age of illness onset (n %) Over median psychiatric treatment days (n %) Over median time spent (%) in hospital (n %) Basic educational level (n %) Pensioned (n %)

Later motor milestone: crude Odds Ratios (95% CI)

Later motor milestone: adjusted Odds Ratios (95% CI) ⁎⁎

− 10 n = 24 18/24 75%

11+ n = 71 26/71 37%

11+ (reference group − 10 months)

11+ (reference group − 10 months)

0.19 (0.07–0.55)

0.15 (0.05–0.46)

18/24 75%

26/69 38%

0.20 (0.07–0.57)1

0.20 (0.07–0.61)

17/24 71%

28/69 41%

0.28 (0.10–0.77)2

0.31 (0.11–0.89)

11/24 46% 17/21 81%

21/71 30% 37/64 58%

0.50 (0.19–1.29) 0.32 (0.10–1.07)

0.31 (0.11–0.94) 0.29 (0.08–1.00)

− 11 n = 27 15/27 56%

12+ n = 68 29/68 43%

12+ (reference group − 11 months) 0.60 (0.24–1.46)

12+ (reference group − 11 months) 0.53 (0.21–1.36)

18/27 67%

26/66 39%

0.33 (0.13–0.83)3

0.36 (0.14–0.95)

19/27 70%

26/66 39%

0.27 (0.11–0.72)4

0.33 (0.12–0.89)

11/27 41% 17/23 74%

21/68 31% 37/62 60%

0.65 (0.26–1.64) 0.52 (0.18–1.51)

0.57 (0.21–1.54) 0.49 (0.17–1.48)

At least 3 words

None or at most 2 words (reference group at least 3 words)

None or at most 2 words (reference group at least 3 words)

n = 50 23/50 46%

None or at most 2 words n = 48 23/48 48%

1.08 (0.49–2.39)

1.05 (0.46–2.42)

25/49 51%

21/47 45%

0.76 (0.35–1.73)

0.92 (0.40–2.13)

27/49 55%

20/47 43%

0.60 (0.27–1.35)

0.77 (0.33–1.79)

16/50 32% 29/44 66%

16/48 33% 27/44 61%

1.06 (0.46–2.47) 0.82 (0.34–1.96)

0.98 (0.40–2.43) 0.79 (0.311.97)

Outcome information ascertained from national registers by the year 2001. Statistically significant values (p b 0.05) are written in bold. Odds Ratios (95% CI) when adjusted for onset age: 10.34 (0.11–1.05), 20.34 (0.12–0.96), 30.35 (0.12–0.99), 40.29 (0.11–0.78). ⁎ Medians for outcome variables: age of illness onset, 22 years; psychiatric treatment days, 269; time spent in hospital (%), 8.22. ⁎⁎ Adjusted for gender, obstetric complications, family history of psychosis, father's social class.

34

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

3. Results Of 109 subjects in the study, 70 (64%) were males. The mean (± SD) age of illness onset was 22.3 years ± 4.1 years (range 14–31). There was information about developmental milestones for 95 individuals with schizophrenia. Of these, 24 (25%) learned to stand unsupported by the age 10 months, 27 (28%) learned to walk by the age 11 months, and 50 (51.0%) spoke at least three words by the age 12 months. (Table 1). The mean IMD score was 0.00 (SD 1.00) and median 0.217 (IQR − 0.602, 0.695). The analyses were made first in those 109 cases detected from the registers and the field study in 1999– 2001 by the year 2001 (Table 2), and additional variables were examined within the subgroup of 59 subjects who participated in the field study and clinical examination in 1999–2001. Compared to those who learned to stand early, individuals with schizophrenia who learned to stand later had statistically significantly later ages of illness onset (Odds Ratio 0.19, 95% Confidence Interval 0.07– 0.55), fewer treatment days in the hospital (OR 0.20, 95% CI 0.07–0.57), and had spent a smaller proportion of time in the hospital after becoming ill (OR 0.28, 95% CI 0.10–0.77). Those who learned to walk later had similarly less treatment days (OR 0.33, 95% CI 0.13– 0.83), and had spent smaller proportion of time in the hospital (OR 0.27, 95% CI 0.11–0.72), compared to those learning earlier. When gender, obstetric complications, family history of psychosis, father's social class and age of illness onset were included in the model most of the associations between age of learning motor milestones and outcome in schizophrenia remained. The associations between developmental milestones and register-based outcome measures are presented in Table 2. Concerning the other, statistically non-significant associations (standing, walking and talking, and outcome variables assessed in the field study in 1999– 2001), the data is not shown. Some statistically significant associations occurred after adjusting. When gender, obstetric complications, family history of psychosis, and father's social class were included into the analyses those who learned to stand later were statistically significantly better educated (OR 0.31, 95% CI 0.11–0.94) compared to early learners. Regarding IMD (not shown in Table 2), higher loading on the IMD component (i.e. indicating delayed motor milestone) was associated with better outcome (i.e. few total days in hospital) (OR 0.61, 95% CI 0.39– 0.94) but this was not statistically significant in the

adjusted models. Similarly, higher loadings on the IMD component was associated with lower proportion of time spent in hospital (0.63, 0.41–0.97), but, after adjusting, the statistically significant association did not remain. Some statistically significant associations occurred in the adjusted models. Higher loadings on the IMD component (i.e. indicating delayed motor milestones) was associated with later age of illness onset (median or over age of illness onset) (0.60, 0.38–0.95), and was associated with better education outcomes (0.55, 0.33–0.92). 4. Discussion 4.1. Main findings While delayed developmental milestones are associated with an increased risk of schizophrenia in the general population, within those with schizophrenia subtle delays in these same milestones do not predict adverse outcomes. Contrary to our hypotheses, individuals with schizophrenia who learned to stand and walk later required significantly fewer days in hospital, and those learning to stand later had also later age of illness onset and were better educated. When considering a continuous summary measure of infant motor development (IMD), later development was associated with later age of illness onset, fewer hospitalizations, and better education. While later attainment of motor milestones appears to increase the risk of schizophrenia, the pattern of association between these milestones and various measures of outcome reveals a more subtle and nuanced pattern: Later psychomotor development is linked to increased risk of schizophrenia but is not a marker and predictor of more severe outcome of illness. When adjusted for potential confounding, most associations between late learning and better outcomes remained. We conclude that the association between later learning (probably mostly within normal variation) and uniformly poor outcome can be rejected, that our study provides strong evidence to weaken the argument that delayed development carries with it a poor prognosis for schizophrenia patients, and that it provides very preliminary evidence for the contrary hypothesis: Namely that subtly delayed development carries with it a better prognosis for schizophrenia patients. 4.2. Theoretical explanation In spite of previous assumptions in the schizophrenia literature that disease risk factors are also prognostic

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

indicators, there are counter-examples to this line of reasoning to be found in somatic illness. For example, while mutations in the BRCA1 and BRCA2 genes substantially increase the risk of developing breast cancer, those same mutations are not predictive of mortality or response to treatment in breast cancer patients, and thus they do not provide useful prognostic information (Rennert et al., 2007). Murray et al. (1992) hypothesized two subtypes of schizophrenia: a neurodevelopmental subtype with genetic risk, obstetric complications, male gender, later developmental milestones, poor school performance and poor course of illness, and non-neurodevelopmental type with less genetic risk and obstetric complications, female gender, relatively normal development and good outcome. Our data, which show later development associated with better outcomes, argue against this classification. In fact, there is evidence that associations between a given independent variable and “outcome” varies substantially according to which measure of outcome in schizophrenia is used. Considering infant motor development as an example, our previous schizophrenia research details stronger associations in schizophrenia than in the general population between later infant development and poorer school performance (Isohanni et al., 2004), and similar strength associations in schizophrenia as in the general population between infant motor development and cognitive function in some domains (Murray et al., 2006). However, infant motor development is less predictive of adult frontocerebellar structure in patients with schizophrenia than it is in the general population (Ridler et al., 2006). Taken together these data and our new results suggest that the relationships between neurodevelopmental markers and adult outcomes in schizophrenia are highly complex. This may reflect the complex multifactorial aetiology of the disorder involving interactions between multiple genes of small effect and multiple environmental risk factors of small effect (McGuffin and Southwick, 2003). 4.3. Methodological discussion Our study has several major strengths. It is based on a large, unselected, geographically representative, general population sample with high rates of follow-up over three decades combined with use of record linkage which minimized the potential selection, information and recall bias. In addition, our diagnostics are accurate; diagnoses for whole sample have been re-checked twice by professionals (Isohanni et al., 1997; Moilanen et al., 2003). Our major results involved standardized child-

35

hood data collected prospectively, including a special examination near the time point that developmental milestones took place. We partly used register-based indicators of outcome. This information can be seen as reliable and the amount of missing information was low. However, register information may be relatively superficial and scarce and relate only minimally to the clinical symptomatology. We have more detailed information about the outcomes for those cases that participated in the field study with personal interviews. Regarding outcome measures in general, there is a serious need for standardized and commonly accepted definition of good and poor outcomes in schizophrenia. This would decrease the large differences in methods of different studies and would also make the comparison on results easier. While our developmental data, being prospectively collected on the basis on an average of 10 visits to child welfare centers per cohort member over a 12 month period, are superior to those used in many schizophrenia studies, there are some limitations. The assessment at age 1 year may have been slightly too early for walking and talking, and a standardized neurological examination was not used. One important limitation of this study is the small sample size. Because of limited statistical power, the negative findings should be taken with caution; a nonsignificant finding does not prove the null hypothesis but instead may reflect the lack of power (Kraemer et al., 2001). These kinds of associations in general are difficult to analyze, since the changes seen in the development of infants who later develop schizophrenia are slight (Isohanni et al., 2001). We did several analyses for the association between age of learning developmental milestones and illness outcome and thus chance findings are possible. However, the consistencies of our findings in different analyses reduce the risk of the findings being due to chance. The cases are still relatively young, and it may be that, in time, different associations between milestones and course of illness emerge. 4.4. Implications for future research Although later development is a weak marker of elevated risk of psychoses, our results concerning its association to the course of illness are thought provoking. While the trajectory of neuromotor development during early life is altered in those who develop schizophrenia, after the onset of illness these same factors appear to relate differently to the course of the

36

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37

illness. Our results show that while delayed developmental milestones are associated with an increased risk of schizophrenia in the general population, delays in these same milestones do not predict adverse outcomes for schizophrenia patients. These findings indicate two important points: The developmental trajectory in schizophrenia is complex, and risk factors for schizophrenia are not necessarily prognostic factors. Role of funding source Funding for this study was provided by the Academy of Finland, Sigrid Juselius Foundation, Stanley Medical Research Institute, and the Finnish Medical Foundation. These sources had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Contributors All the authors had contributed to the designing of the study. E Jääskeläinen managed the literature searches and analyses. E Jääskeläinen and J Miettunen undertook the statistical analysis, and author E Jääskeläinen wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest E Jääskeläinen has received unrestricted grants from: Oy H. Lundbeck Ab Finland, the University of Oulu, the Research Foundation of Orion Pharma Corporation, Medical Doctors' Association Duodecim, The Finnish Medical Foundation, Gates European Trust, GlaxoSmithKline. E Jääskeläinen does not have any financial ties to pharmaceutical companies. J Miettunen has no conflict of interest. J Veijola has no conflict of interest. J McGrath has no conflict of interest. G Murray has no conflict of interest. P Jones has acted as consultant for BMS & Otsuka, and Eli Lilly & Co. M Isohanni has acted in scientific advisory board of AstraZeneca and Janssen-Cilag. Acknowledgements This work was supported by the unrestricted grants from the Academy of Finland, Sigrid Juselius Foundation, Stanley Medical Research Institute, and the Finnish Medical Foundation.

References Alaräisänen, A., Miettunen, J., Lauronen, E., Räsänen, P., Isohanni, M., 2006. Good school performance is a risk factor of suicide in psychoses but a protective factor in non-psychotic population. A 35-year follow-up of the Northern Finland 1966 Birth Cohort. Acta Psychiatr. Scand. 114, 357–362. Cannon, M., Caspi, A., Moffitt, T.E., Harrington, H., Taylor, A., Murray, R.M., Poulton, R., 2002. Evidence for early-childhood, pan-developmental impairment specific to schizophreniform disorder: results from a longitudinal birth cohort. Arch. Gen. Psychiatry 59, 449–456. Castle, D.J., Wessely, S., Murray, R.M., 1993. Sex and schizophrenia: effects of diagnostic stringency, and associations with and premorbid variables. Br. J. Psychiatry 162, 658–664.

Fish, B., Marcus, J., Hans, S.l., Auerbach, J.G., Perdue, S., 1992. Infants at risk for schizophrenia: sequelae of a genetic neurointegrative defect. A review and replication analysis of pandysmaturation in the Jerusalem Infant Development Study. Arch. Gen. Psychiatry 49, 221–235. Harrison, G., Hopper, K., Craig, T., Laska, E., Siegel, C., Wanderling, J., Dube, K.C., Ganev, K., Giel, R., an der Heiden, W., Holmberg, S.K., Janca, A., Lee, P.W.H., Leon, C.A., Malhotra, S., Marsella, A.J., Nakane, Y., Sartorius, N., Shen, Y., Skoda, R., Thara, R., Tsirkin, S.J., Varma, V.K., Walsh, D., Wiersma, D., 2001. Recovery from psychotic illness: a 15- and 25-year international follow-up study. Br. J. Psychiatry 178, 506–517. Isohanni, M., Mäkikyrö, T., Moring, J., Räsänen, P., Hakko, H., Partanen, U., Koiranen, M., Jones, P., 1997. Comparison of clinical and research DSM-III-R diagnoses of schizophrenia in a Finnish national birth cohort. Soc. Psychiatry Psychiatr. Epidemiol. 32, 303–308. Isohanni, M., Jones, P., Moilanen, K., Veijola, J., Oja, H., Koiranen, M., Jokelainen, J., Croudace, T., Järvelin, M.-R., 2001. Early developmental milestones in adult schizophrenia and other psychoses. A 31-year follow-up of the Northern Finland 1966 Birth Cohort. Schizophr. Res. 52, 1–19. Isohanni, M., Murray, G.K., Jokelainen, J., Croudace, T., Jones, P.B., 2004. The persistence of developmental markers in childhood and adolescence and risk for schizophrenic psychoses in adult life. A 34-year follow-up of the Northern Finland 1966 Birth Cohort. Schizophr. Res. 71, 213–225. Johnstone, E.C., Frith, C.D., Lang, F.H., Owens, D.G., 1995. Determinants of the extremes of outcome in schizophrenia. Br. J. Psychiatry 167 (5), 604–609. Jones, P.B., Rodgers, B., Murray, R., Marmot, M., 1994. Child developmental risk factors for adult schizophrenia in the British 1946 Birth Cohort. Lancet 344, 1398–1402. Jones, P.B., Rantakallio, P., Hartikainen, A.-L., Isohanni, M., Sipilä, P., 1998. Schizophrenia as a long-term outcome of pregnancy, delivery, and perinatal complications: a 28-year follow-up of the 1966 North Finland General Population Birth Cohort. Am. J. Psychiatry 155, 355–364. Jonsson, H., Nyman, A.K., 1984. Prediction of outcome in schizophrenia. Acta Psychiatr. Scand. 69, 274–291. Jonsson, H., Nyman, A.K., 1991. Predicting long-term outcome in schizophrenia. Acta Psychiatr. Scand. 83, 342–346. Kraemer, H.C., Stice, E., Kadzin, A., Offord, D., Kupfer, D., 2001. How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. Am. J. Psychiatry 158, 848–856. Maxwell, M.E., 1992. Family Interview for Genetic Studies (FIGS): Manual for the FIGS. Clinical Neurogenetics Branch. National Institute Of Mental Health, Bethesda, MD. Intramural Research Program. McGuffin, P., Southwick, L., 2003. Fifty years of the double helix and its impact on psychiatry. Aust. N. Z. J. Psychiatry 37, 657–661. Moilanen, K., Veijola, J., Läksy, K., Mäkikyrö, T., Miettunen, J., Kantojärvi, L., Kokkonen, P., Karvonen, J., Herva, A., Joukamaa, M., Järvelin, M.-R., Moring, J., Jones, P., Isohanni, M., 2003. Reasons for the diagnostic discordance between clinicians and researchers in schizophrenia in the Northern Finland 1966 Birth Cohort. Soc. Psychiatry Psychiatr. Epidemiol. 38, 305–310. Murray, R.M., O'Callaqhan, E., Castle, D.J., Lewis, S.W., 1992. A neurodevelopmental approach to the classification of schizophrenia. Schizophr. Bull. 18 (2), 319–332. Murray, G.K., Jones, P.B., Moilanen, K., Veijola, J., Miettunen, J., Cannon, T.D., Isohanni, M., 2006. Infant motor development and

E. Jääskeläinen et al. / Schizophrenia Research 99 (2008) 29–37 adult cognitive functions in schizophrenia. Schizophr. Res. 81 (1), 65–74. Rantakallio, P., 1969. Groups at risk in low birth weight infants and perinatal mortality. Acta Paediatr. Scand. 193 (Suppl.), 1–71. Rantakallio, P., Von Wendt, L., Mäkinen, H., 1985. Influence of social background on psychomotor development in the first year of life and its correlation with later intellectual capacity: a prospective cohort study. Early Hum. Dev. 11, 141–148. Räsänen, S., Veijola, J., Hakko, H., Joukamaa, M., Isohanni, M., 1999. Gender differences in incidence and age at onset of DSM-III-R schizophrenia. Preliminary results of the Northern Finland 1966 Birth Cohort Study. Schizophr. Res. 37, 197–198. Rennert, G., Bisland-Naggan, S., Barnett-Griness, O., Bar-Joseph, N., Zhang, S., Rennert, H.S., Narod, S.A., 2007. Clinical outcomes of breast cancer in carriers or BRCS1 and BRCA2 mutations. N. Engl. J. Med. 357, 115–123. Ridler, K., Veijola, J., Tanskanen, P., Haapea, M., Miettunen, J., Pyhtinen, J., Suckling, J., Murray, G., Jones, P.B., Isohanni, M., Bullmore, E.T., 2006. Fronto-cerebellar systems are associated with infant motor and adult executive functions in healthy adults but not in schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 17:103 (42), 15651–15656. Rossi, A., Pollice, R., Daneluzzo, E., Marinangeli, M.G., Stratta, P., 2000. Behavioral neurodevelopmental abnormalities and schizophrenic disorder: a retrospective evaluation with the Childhood Behavior Checklist (CBCL). Schizophr. Res. 44, 121–128. Sintonen, H., 2001. The 15D instrument of health-related quality of life: properties and applications. Ann. Med. 33, 328–336.

37

Spitzer, R.L., Williams, J.B.W., Gibbon, M., First, M.B., 1989. Structural Diagnostic Interview For DSM-III-R. Structured Clinical Interview For DSM-III-R — patient edition (Scid-P, 9/1/ 89 Version). Biometrics Research Department, New York State Psychiatric Institute, New York, Ny. Spitzer, R.L., Gibbon, M., Endicott, J., 2000. Global Assessment Scale (GAS), Global Assessment of Functioning (GAF) Scale, Social and Occupational Functioning Assessment Scale (SOFAS). American Psychiatric Association. Handbook of Psychiatric Measures. American Psychiatric Association, Washington. Suvisaari, J.M., Haukka, J., Tanskanen, A., Lönnqvist, J.K., 1998. Age at onset and outcome in schizophrenia are related to the degree of familial loading. Br. J. Psychiatry 173 (12), 494–500. Tanskanen, P., Veijola, J.M., Piippo, U.K., Haapea, M., Miettunen, J., Pyhtinen, J., Bullmore, E.T., Jones, P.B., Isohanni, M., 2005. Hippocampus and amygdala volumes in schizophrenia and other psychoses in the Northern Finland 1966 Birth Cohort. Schizophr. Res. 75 (2–3), 283–294. Tuori, T., Lehtinen, V., Hakkarainen, A., Jääskeläinen, J., Kokkola, A., Ojanen, M., Pylkkänen, K., Salokangas, R.K.R., Solantaus, J., Alanen, Y.O., 1998. The Finnish National Schizophrenia Project 1981–1987: 10-year evaluation of its results. Acta Psychiatr. Scand. 97, 10–17. van Os, J., Takei, N., Castle, D.J., Wessely, S., Der, G., Murray, R.M., 1995. Premorbid abnormalities in mania, schizomania, acute schizophrenia and chronic schizophrenia. Soc. Psychiatry Psychiatr. Epidemiol. 30 (6), 274–278.