| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From The Centre for Mental Health Research (A.F.J.), Australian National University, Canberra; Department of Psychology (M.P., A.S., D.S.), Royal Children’s Hospital/University of Melbourne, Melbourne; and the Centre for Bioinformation Science and John Curtin School of Medical Research (Y.Z., S.E.), Australian National University, Canberra, Australia.
Address reprint requests to: Anthony F. Jorm, PhD, DSc, Centre for Mental Health Research, Australian National University, Canberra, ACT 0200, Australia. Email: Anthony.Jorm{at}anu.edu.au
Received for publication May 31, 2002; revision received October 7, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: DNA was collected from 683 Australian children who had participated in a longitudinal study of childhood temperament from 4 to 8 months up to 17 to 18 years. Associations were examined between APOE genotype and a range of measures of activity and hyperactivity.
RESULTS: No associations were found.
CONCLUSIONS: The earlier Finnish finding could not be replicated despite adequate statistical power.
Key Words: apolipoprotein E, • temperament, • children, • activity.
Abbreviations: APOE = apolipoprotein E;; EAS = Emotionality Activity and Sociability;; PCR = polymerase chain reaction;; RBPC = Revised Behavior Problem Checklist;; SATI = School Aged Temperament Inventory;; STSI = Short Temperament Scale for Infants;; STST = Short Temperament Scale for Toddlers.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
2, 3 and 4 alleles of the APOE gene which is on chromosome 19. In Caucasian populations, the 3 allele is by far the most common (
78%), followed by the 4 (14%) and the 2 (8%) alleles (1).
The APOE genotype has been extensively investigated for associations with human disease. Relative to the common 3 allele, the 4 allele is known to increase risk for Alzheimer’s disease and the 2 allele to decrease risk (1, 2). There have also been numerous investigations of associations with vascular disease. The 4 allele is associated with higher levels of low density lipoprotein and total cholesterol and the 2 allele with lower levels. Furthermore, the 4 allele is associated with an increased risk of coronary heart disease (3) and ischemic cerebrovascular disease (4), but the 2 allele does not appear to be protective for either disease.
Given the association between APOE genotype and coronary heart disease, Keltikangas-Järvinen and colleagues have hypothesized that this genotype might also be associated with behavioral risk factors for this disease (5). According to this hypothesis, the behaviors do not in themselves increase risk, but are an indicator of a constitutional weakness that increases risk of atherosclerosis. To test the hypothesis, APOE phenotype was investigated for associations with Type A behavior and with the temperamental traits of activity, sociability, and emotionality in 1,577 healthy Finnish 3 to 18 year olds who were followed up over a period of 9 years. E4 phenotype was found to be associated with increased activity. This association manifested somewhat differently with age, being seen with motor activity in childhood, but with "mental vitality" in adolescence and young adulthood. The same study found no consistent associations of the E4 isoform with either Type A behavior or the temperamental traits of sociability and emotionality.
Although the APOE genotype has been the focus of considerable research over the past decade, it is surprising that there have been no attempts to replicate the finding of an association with activity. We therefore examined the association of the APOE genotype with temperamental traits involving increased activity using data from an Australian longitudinal study, the Australian Temperament Project, spanning from infancy through the late teens (6). The major aim of the Project is to investigate the influence of a child’s temperament on health and development in both the short and longer term, and the ways in which it affects pathways to adjustment and maladjustment. This study is one of a series from the Australian Temperament Project reporting on the association of genetic polymorphisms with temperament and behavior problems (7–10).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Questionnaire
Behavioral data were collected by postal survey. Mothers completed the STSI (11) at child age 4 to 8 months, the STST (12) at 1 to 2 and 2 to 3 years, the EAS Temperament Questionnaire (13) at 9 to 10 years, and the SATI (14) at 11 to 12 through to 15 to 16 years. Further details on the content of the temperament dimensions can be found in Sanson et al. (15). To assess behavior problems, mothers completed the Behar Pre-school Behavior Questionnaire (16) at 3 to 4 years, the Rutter Problem Behavior Questionnaire (17) at 5 to 6 through to 12 to 13 years, and the RBPC (18) at 13 to 14 to 17 to 18 years. These instruments have published reliability and validity data (19). Teachers also completed the Rutter questionnaire at ages 5 to 6, 7 to 8, and 11 to 12 years. This covers parallel behavior problem domains to the parent questionnaire. The children self-completed parallel forms of the Rutter questionnaire at ages 11 to 12 and 12 to 13 and the RBPC at 13 to 14 to 17 to 18 years, adapted for child report by the authors.
In the present analyses, we used temperamental measures associated with higher levels of activity: Activity-Reactivity, Distractibility and Activity. We also analyzed the Hyperactive scales from the Behar and Rutter questionnaires and the Attention and Motor Tension Excess scales of the RBPC.
Procedure
Families who agreed to participate received a home visit from a research assistant who took cheek swabs from the child. The cheek swab involved brushing the inside of the cheek with a cotton bud for 30 seconds, with two cheek swabs being taken on each child. The study was approved by the ethics committee of the University of Melbourne and conformed to the principles of the Declaration of Helsinki.
Genotyping
DNA was extracted by a modification of the method of Richards et al. (20). APOE genotypes were determined by PCR amplification of a 234 base-pair fragment of exon 4 of the APOE gene followed by digestion with CfoI. The APOE 2, 3, and 4 alleles have distinct combinations of CfoI (or HhaI) restriction fragments. These were separated by polyacrylamide gel electrophoresis and displayed by silver staining.
Statistical analysis
Differences in mean scale scores between the six genotypes were tested using analysis of variance, followed by post hoc Scheffé tests where this was significant. Because the numbers in the some of the genotype groups were small, we also combined the genotypes containing an 4 allele and compared them to all other genotypes combined. The p < .05 significance level was used.
The nQuery software (21) was used for power calculations, using a two-tailed 
value of .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
2/2 1%, 2/3 12%, 2/4 3%, 3/3 59%, 3/4 23% and 4/4 2%. The allele frequencies were 2 9% (95% CI: 7–10), 3 77% (95% CI: 75–79), and 4 14% (95% CI: 13–16). These frequencies are very close to those reported in the literature for Caucasians (1). However, compared with the Finnish sample previously investigated (5), the present sample had a higher frequency of the 2 allele (9% vs. 4%) and a lower frequency of the 4 allele (14% vs. 19%).
Analyses of variance on the 29 dependent variables found only one significant effect. At age 13 to 14 years, the SATI activity temperament factor (parent report) differed between genotypes, with means (SDs) of 3.26 (0.73) for 2/2, 2.51 (0.71) for 2/3, 2.91 (0.47) for 2/4, 2.62 (0.73) for 3/3, 2.55 (0.66) for 3/4, and 3.00 (0.80) for 4/4. However, post hoc tests found no significant differences between genotypes in this case. The data were also analyzed comparing genotypes with one or more 4 alleles vs. all others combined. No significant differences were found.
It is possible that the present study lacked the statistical power to detect the effect reported in the Finnish study. To estimate the statistical power of the present study, it was assumed that the significant results reported for activity variables in the Finnish study are the true population effects. For simplicity, these effects were calculated for genotypes with the 4 allele vs. those without. The power to detect the significant effects reported for activity variables in the Finnish study ranged from 74% to 92% (two-tailed = 0.05) at most ages, and from 57% to 79% at the two ages where only two-thirds of the sample was assessed.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
4 allele was associated with increased activity (5). It may be that the Finnish finding was a Type I error, but other explanations are possible and these are discussed below. The first is that the present study lacked statistical power. However, a power analysis showed that this is unlikely to be the reason for the negative results.
Another potential source of the differing findings is the measures of temperament used. The Finnish study used measures that the authors developed for the purpose and these did not have high reliability. Our measures were more standard and data on reliability and validity have been reported (6, 22, 23). It is not known how they would correlate with the measures used in the Finnish study.
Another difference between the studies is in the ethnicity of the Finnish and Australian Caucasian samples. The Finns are a population isolate, whereas the Australian Caucasian population is mixed. Compared with the Australians, the Finnish sample had a higher frequency of 4 and a lower frequency of 2, which is consistent with other research comparing Finns with other European samples (24). The strength of the association between APOE and Alzheimer’s disease is known to vary across ethnic groups (1), so it is possible that any association with temperament could also vary.
In conclusion, the present study has considerable strengths. It involved validated measures of temperament and behavior problems on a large general population sample and it has longitudinal data covering the whole of childhood and adolescence. Nevertheless, this study was unable to replicate the results of the earlier Finnish study and these must be regarded as unreliable in the absence of further confirmation.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. M. Teasdale, G. D. Murray, and J. A. R. Nicoll The association between APOE {varepsilon}4, age and outcome after head injury: a prospective cohort study Brain, November 1, 2005; 128(11): 2556 - 2561. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
