bs_bs_banner
doi: 10.1111/ppe.12165
41
Does Late Childbearing Increase the Risk for Behavioural Problems in Children? A Longitudinal Cohort Study Jessica E. Tearne,a,b Monique Robinson,a Peter Jacoby,a Jianghong Li,a,c,d John Newnham,e Neil McLeanb a
Telethon Kids Institute, The University of Western Australia, Perth, Australia b
School of Psychology, The University of Western Australia, Perth, Australia c
WZB Berlin Social Research Center, Berlin, Germany
d
Centre for Population Health Research, Faculty of Health Sciences, Curtin University, Perth, Western Australia, Australia
e
School of Women’s and Infants’ Health, King Edward Memorial Hospital, The University of Western Australia, Perth, Australia
Abstract Background: This study aimed to examine the relationship between advanced parental age and behavioural outcomes in offspring in a longitudinal cohort of children in Western Australia. Methods: The Western Australian Pregnancy Cohort (Raine) is a prospective study of 2900 pregnancies. Offspring were followed up at ages 2, 5, 8, 10, 14, and 17 years, and 1754 adolescents were available for follow-up at 17 years. The Child Behaviour Checklist was used to measure child behaviour, including internalising (e.g. anxious/withdrawn) and externalising (e.g. aggressive/destructive) behaviours. Results: There was a significant linear relationship between maternal age and total internalising and externalising behaviour outcomes, but not paternal age. Increasing maternal age was associated with decreasing risk for problem behaviours in offspring. In the categorical models, young maternal age (20–24 years) was associated with significantly increased risk for problem behaviours in offspring relative to offspring of parents in the reference group (25–29 years), and a paternal age of 35–39 years was associated with decreased risk for total behaviour morbidity in offspring. Conclusions: This study showed no evidence that late fatherhood is associated with adverse behavioural outcomes in offspring. Increasing maternal age was found to be a protective factor for child behaviour morbidity. Keywords: Raine study, CBCL, behavioural development, mental health, late parenthood.
The postponement of childbirth in developed countries has become a significant and widespread phenomenon across many Western countries.1 Increasingly, parental age at childbirth is being recognised as an important predictor of psychiatric outcomes in offspring. There is evidence to suggest that advanced paternal age may be involved in the aetiology of schizophrenia,2,3 bipolar disorder,4,5 and autism spectrum disorders.6,7 Buizer-Voskamp et al.8 also found a ‘U’-shaped relationship between paternal age and major depressive disorder in offspring, with increased risk for children of both younger and older fathers. Where efforts have been made to link maternal age to specific psychiatric outcomes, this relationship tends to become non-significant once paternal age is accounted for. Menezes et al.5 found an associaCorrespondence: Jessica E. Tearne, School of Psychology M304, 35 Stirling Hwy, Crawley, WA 6009, Australia. E-mail: [email protected]
© 2014 John Wiley & Sons Ltd Paediatric and Perinatal Epidemiology, 2015, 29, 41–49
tion between advanced maternal age and increased risk of bipolar disorder in offspring, but this relationship was not found in another study.4 Advanced maternal age has also been linked to increased risk for autism spectrum disorders,9–11 but not in all studies.7,12,13 These findings relate to more serious psychiatric diagnoses, but comparatively little is known about the potential impact of parental age on more common behavioural problems in offspring. Several studies have suggested increased risk for offspring behaviour problems associated with younger motherhood, particularly teen motherhood, in younger and teenaged children.14–17 Harden et al. suggested that relative to their later born siblings, children of teenage mothers were at increased risk for internalising problems, substance misuse, and internalising problems, and that this relationship was independent of family background factors that may confound the relationship.17 Other studies have suggested that maternal age is not
42
J. E. Tearne et al.
a causal factor for behaviour problems, but rather that the relationship between maternal age and behaviour problems can be accounted for by the disadvantage often associated with becoming a mother at an earlier age (e.g. increased likelihood of low socioeconomic status, non-completion of schooling).15,18 Few studies have specifically examined the effect of late parenthood (maternal and paternal age at birth) on offsprings’ behavioural outcomes. Saha et al.19 reported a decreased risk of externalising behaviours in 7-year-old offspring (such as aggression and conduct problems) for every 5-year increase in maternal age, but an increased risk of internalising behaviours (such as anxiety, withdrawal, depression) for every 5-year increase in maternal age. They also found that offspring (age 7) of older fathers were more likely to show adverse externalising behaviours, with every 5-year increase in paternal age corresponding to a 12 per cent increase in risk of higher scores on measures of externalising behaviours. However, the data for this study were collected between 1959 and 1965, and the reasons for delaying parenthood at this time may not be the same three decades later. In a more recent sample, D’Onofrio et al.16 found that for every 1 year increase in maternal age at childbirth, there was a significant reduction in the risk of conduct problems and delinquent behaviours. Another study identified a ‘U’-shaped relationship between parental age and social functioning in 16–17-year-old males (e.g. number of friends, sociability, and intimate relationships) with offspring of both mothers and fathers under the age of 20 or over the age of 40 more likely to display poor social functioning.20 Fergusson and Woodward21 found that increased maternal age was associated with decreased risk of juvenile crime, substance misuse, and mental health problems in offspring. The relationships were largely explained by associations between maternal age and child-rearing practices and home environments. While there is an accumulating literature on the relationship between parental age and severe psychopathology in offspring, relatively little attention has been paid to childbearing in the late 30s and early 40s on a broader spectrum of behaviours through childhood and adolescence. Given the trend towards delayed childbearing across Western nations, investigation of the relationship between advanced maternal and paternal age and behavioural outcomes in offspring is warranted in order to understand the implications of this trend for society and individual families. In this
study, we aimed to establish what links, if any, exist between advanced parental age and internalising and externalising behaviours in offspring through childhood and adolescence. We also examined whether these effects persisted once we accounted for various psychosocial and economic factors known to be associated with advanced parental age.
Methods Study population The Western Australian Pregnancy Cohort (Raine) is a prospective population-based pregnancy cohort study of 2868 livebirths followed from 16 weeks gestation. Pregnant women between 16 and 20 weeks gestation (n = 2900) were recruited between May 1989 and November 1991 through the public antenatal clinic at King Edward Memorial Hospital (KEMH) and nearby private clinics in Perth, Western Australia. To be eligible, women were required to have sufficient English language skills, an expectation to deliver at KEMH, and an intention to reside in Western Australia to allow for follow-up of their child.22 At total of 2868, infants were available at birth for follow-up. Information on recruitment and study design are provided in detail elsewhere.22 Data regarding psychosocial and demographic characteristics were provided at enrolment and updated with further data collection at 34 weeks gestation. Families were followed up at ages 1, 2, 3, 5, 8, 10, 14, 17, 20, and 23 years using questionnaires and physical measurements. This research was approved by the Human Research Ethics Committees at KEMH and Princess Margaret Hospital for Children in Perth, Western Australia. Written parental and adolescent consent was obtained at recruitment and at each follow-up.
Attrition Rates of attrition in the Raine Study have been described elsewhere.23 At the 17-year follow-up, 2352 adolescents and their families were eligible for follow-up (480 had withdrawn and 36 were deceased). Of those eligible to participate, 414 declined to participate and 184 were unable to be traced, leaving 1754 adolescents and their families who were available to participate in the 17-year follow-up. This represented 75% of those eligible to © 2014 John Wiley & Sons Ltd Paediatric and Perinatal Epidemiology, 2015, 29, 41–49
Parental age and offspring behaviour participate at age 17 and 61% of the original cohort of 2868 livebirths. Further information on the numbers of participants in each follow-up is provided in Table S1. Characteristics of participants and missing participants are presented in Table S2a–b.
Outcome variables Child Behaviour Checklist (CBCL) Data on child behaviour were collected at the 2-, 5-, 8-, 10, 14-, and 17-year follow-ups. The CBCL for ages 4–18 (CBCL/4–18) is a 118-item empirically validated checklist of problem behaviours based on the previous 6 months which was completed by a parent/guardian of children at ages 5, 8, 10, 14, and 17. The CBCL for ages 2–3 (CBCL/2–3) is a 99 item empirically validated checklist of specific problem behaviours derived from the CBCL/4–18. This was completed by a parent/ guardian at the 2-year follow-up. The CBCL is used to measure behaviour as a continuous score, in addition to apportioning factors into clinical syndrome scales including, withdrawal, anxious/depressed, somatic, social, attention, thought, aggression, and delinquency scales.24 Both the CBCL/2–3 and the CBCL/4–18 produce raw scores that can be transformed into summary T scores for total behaviour, internalising behaviour (withdrawal, anxious/depressed, and somatic), and externalising behaviours (delinquency and aggression). The recommended clinical cut-off scores (T ≥ 60) were applied to the CBCL T scores to obtain three binary variables indicative of clinically significant scores on the total, internalising, and externalising problem scales.24 In this study, scoring above this clinical cut-off is referred to as behavioural morbidity.
Predictor variables Maternal and paternal age was considered as both continuous and categorical variables. For the categorical models, maternal and paternal age at birth of the study child was classified within 5-year age categories as per those widely used in population fertility data:1 $24 000 AUD per annum/ ≤$24 000 AUD per annum; threshold representing the poverty line at the time26), and gender of offspring.
Statistical analyses Logistic regression models with generalised estimating equations were used to generate odds ratios which estimate the effect of maternal and paternal age on clinically significant T scores in offspring from age 2 to 17 years. Odds ratios when parental age was included as a categorical measure were generated relative to a reference category of parents aged 25– 29 years. An exchangeable correlation matrix was selected for the models. These models were used to account for repeated observations of the same individuals over time. The first model examined maternal and paternal age and offspring behavioural outcome while accounting for age of other parent and the confounding variables listed above. We selected confounding variables known to be associated with child behaviour outcomes, and all variables were entered into the final adjusted model. We then ran the model selecting only for first-time mothers. Ostensibly, the psychosocial profile of a mother of advanced age having her first baby is quite different to that of a mother of advanced age having another baby, which may impact upon the relationship of interest. The
44
J. E. Tearne et al. gender all accounted for a significant proportion of the variance in all three models (P = 0.001) such that offspring who were male, had mothers who did not finish high school, had mothers who smoked during pregnancy, or were born to families living below the poverty line were at significantly increased risk of clinical levels of behaviour problems. Maternal ethnicity, parity, and alcohol use during pregnancy were not significantly associated with behaviour outcomes. Correlations between raw CBCL scores at each follow up are presented in Table 2.
results of an unadjusted model examining only the relationship between maternal and paternal age and offspring outcome with adjustment for age of other parent are presented in Table S2. IBM SPSS Statistics 19.0 was used for the analyses.
Results Frequency data for each of the variables by mothers’ age are presented in Table 1. Behavioural problems (CBCL total/internalising/externalising T-score ≥ 60; population averaged estimates of T-scores observed over each follow-up), maternal education, maternal ethnicity, smoking in pregnancy, and total family income were all significantly related to both maternal and paternal age. Maternal and paternal age were strongly correlated with each other, r = 0.61. Preliminary analyses indicated that there was not a significant interaction between mothers’ and father’s age and behavioural outcomes, and this interaction was not retained in further analyses. Maternal education, total family income, maternal smoking in pregnancy, and
Linear models When used as a continuous variable, there was a significant linear relationship between maternal age and total, internalising, and externalising behaviour morbidity in offspring (Table 3). Increasing maternal age was associated with a decreasing risk for problem behaviours in offspring. No significant relationships between paternal age and behaviour outcomes were noted.
Table 1. Characteristics of participants by mothers’ age Maternal age (years)
Pa Behavioural problems (% with T score ≥ 60) Total Internalising Externalising Maternal education at pregnancy 12 or more years education 24 000 per annum ≤24 000 per annum Gender of child Male Female a
doi: 10.1111/ppe.12165
41
Does Late Childbearing Increase the Risk for Behavioural Problems in Children? A Longitudinal Cohort Study Jessica E. Tearne,a,b Monique Robinson,a Peter Jacoby,a Jianghong Li,a,c,d John Newnham,e Neil McLeanb a
Telethon Kids Institute, The University of Western Australia, Perth, Australia b
School of Psychology, The University of Western Australia, Perth, Australia c
WZB Berlin Social Research Center, Berlin, Germany
d
Centre for Population Health Research, Faculty of Health Sciences, Curtin University, Perth, Western Australia, Australia
e
School of Women’s and Infants’ Health, King Edward Memorial Hospital, The University of Western Australia, Perth, Australia
Abstract Background: This study aimed to examine the relationship between advanced parental age and behavioural outcomes in offspring in a longitudinal cohort of children in Western Australia. Methods: The Western Australian Pregnancy Cohort (Raine) is a prospective study of 2900 pregnancies. Offspring were followed up at ages 2, 5, 8, 10, 14, and 17 years, and 1754 adolescents were available for follow-up at 17 years. The Child Behaviour Checklist was used to measure child behaviour, including internalising (e.g. anxious/withdrawn) and externalising (e.g. aggressive/destructive) behaviours. Results: There was a significant linear relationship between maternal age and total internalising and externalising behaviour outcomes, but not paternal age. Increasing maternal age was associated with decreasing risk for problem behaviours in offspring. In the categorical models, young maternal age (20–24 years) was associated with significantly increased risk for problem behaviours in offspring relative to offspring of parents in the reference group (25–29 years), and a paternal age of 35–39 years was associated with decreased risk for total behaviour morbidity in offspring. Conclusions: This study showed no evidence that late fatherhood is associated with adverse behavioural outcomes in offspring. Increasing maternal age was found to be a protective factor for child behaviour morbidity. Keywords: Raine study, CBCL, behavioural development, mental health, late parenthood.
The postponement of childbirth in developed countries has become a significant and widespread phenomenon across many Western countries.1 Increasingly, parental age at childbirth is being recognised as an important predictor of psychiatric outcomes in offspring. There is evidence to suggest that advanced paternal age may be involved in the aetiology of schizophrenia,2,3 bipolar disorder,4,5 and autism spectrum disorders.6,7 Buizer-Voskamp et al.8 also found a ‘U’-shaped relationship between paternal age and major depressive disorder in offspring, with increased risk for children of both younger and older fathers. Where efforts have been made to link maternal age to specific psychiatric outcomes, this relationship tends to become non-significant once paternal age is accounted for. Menezes et al.5 found an associaCorrespondence: Jessica E. Tearne, School of Psychology M304, 35 Stirling Hwy, Crawley, WA 6009, Australia. E-mail: [email protected]
© 2014 John Wiley & Sons Ltd Paediatric and Perinatal Epidemiology, 2015, 29, 41–49
tion between advanced maternal age and increased risk of bipolar disorder in offspring, but this relationship was not found in another study.4 Advanced maternal age has also been linked to increased risk for autism spectrum disorders,9–11 but not in all studies.7,12,13 These findings relate to more serious psychiatric diagnoses, but comparatively little is known about the potential impact of parental age on more common behavioural problems in offspring. Several studies have suggested increased risk for offspring behaviour problems associated with younger motherhood, particularly teen motherhood, in younger and teenaged children.14–17 Harden et al. suggested that relative to their later born siblings, children of teenage mothers were at increased risk for internalising problems, substance misuse, and internalising problems, and that this relationship was independent of family background factors that may confound the relationship.17 Other studies have suggested that maternal age is not
42
J. E. Tearne et al.
a causal factor for behaviour problems, but rather that the relationship between maternal age and behaviour problems can be accounted for by the disadvantage often associated with becoming a mother at an earlier age (e.g. increased likelihood of low socioeconomic status, non-completion of schooling).15,18 Few studies have specifically examined the effect of late parenthood (maternal and paternal age at birth) on offsprings’ behavioural outcomes. Saha et al.19 reported a decreased risk of externalising behaviours in 7-year-old offspring (such as aggression and conduct problems) for every 5-year increase in maternal age, but an increased risk of internalising behaviours (such as anxiety, withdrawal, depression) for every 5-year increase in maternal age. They also found that offspring (age 7) of older fathers were more likely to show adverse externalising behaviours, with every 5-year increase in paternal age corresponding to a 12 per cent increase in risk of higher scores on measures of externalising behaviours. However, the data for this study were collected between 1959 and 1965, and the reasons for delaying parenthood at this time may not be the same three decades later. In a more recent sample, D’Onofrio et al.16 found that for every 1 year increase in maternal age at childbirth, there was a significant reduction in the risk of conduct problems and delinquent behaviours. Another study identified a ‘U’-shaped relationship between parental age and social functioning in 16–17-year-old males (e.g. number of friends, sociability, and intimate relationships) with offspring of both mothers and fathers under the age of 20 or over the age of 40 more likely to display poor social functioning.20 Fergusson and Woodward21 found that increased maternal age was associated with decreased risk of juvenile crime, substance misuse, and mental health problems in offspring. The relationships were largely explained by associations between maternal age and child-rearing practices and home environments. While there is an accumulating literature on the relationship between parental age and severe psychopathology in offspring, relatively little attention has been paid to childbearing in the late 30s and early 40s on a broader spectrum of behaviours through childhood and adolescence. Given the trend towards delayed childbearing across Western nations, investigation of the relationship between advanced maternal and paternal age and behavioural outcomes in offspring is warranted in order to understand the implications of this trend for society and individual families. In this
study, we aimed to establish what links, if any, exist between advanced parental age and internalising and externalising behaviours in offspring through childhood and adolescence. We also examined whether these effects persisted once we accounted for various psychosocial and economic factors known to be associated with advanced parental age.
Methods Study population The Western Australian Pregnancy Cohort (Raine) is a prospective population-based pregnancy cohort study of 2868 livebirths followed from 16 weeks gestation. Pregnant women between 16 and 20 weeks gestation (n = 2900) were recruited between May 1989 and November 1991 through the public antenatal clinic at King Edward Memorial Hospital (KEMH) and nearby private clinics in Perth, Western Australia. To be eligible, women were required to have sufficient English language skills, an expectation to deliver at KEMH, and an intention to reside in Western Australia to allow for follow-up of their child.22 At total of 2868, infants were available at birth for follow-up. Information on recruitment and study design are provided in detail elsewhere.22 Data regarding psychosocial and demographic characteristics were provided at enrolment and updated with further data collection at 34 weeks gestation. Families were followed up at ages 1, 2, 3, 5, 8, 10, 14, 17, 20, and 23 years using questionnaires and physical measurements. This research was approved by the Human Research Ethics Committees at KEMH and Princess Margaret Hospital for Children in Perth, Western Australia. Written parental and adolescent consent was obtained at recruitment and at each follow-up.
Attrition Rates of attrition in the Raine Study have been described elsewhere.23 At the 17-year follow-up, 2352 adolescents and their families were eligible for follow-up (480 had withdrawn and 36 were deceased). Of those eligible to participate, 414 declined to participate and 184 were unable to be traced, leaving 1754 adolescents and their families who were available to participate in the 17-year follow-up. This represented 75% of those eligible to © 2014 John Wiley & Sons Ltd Paediatric and Perinatal Epidemiology, 2015, 29, 41–49
Parental age and offspring behaviour participate at age 17 and 61% of the original cohort of 2868 livebirths. Further information on the numbers of participants in each follow-up is provided in Table S1. Characteristics of participants and missing participants are presented in Table S2a–b.
Outcome variables Child Behaviour Checklist (CBCL) Data on child behaviour were collected at the 2-, 5-, 8-, 10, 14-, and 17-year follow-ups. The CBCL for ages 4–18 (CBCL/4–18) is a 118-item empirically validated checklist of problem behaviours based on the previous 6 months which was completed by a parent/guardian of children at ages 5, 8, 10, 14, and 17. The CBCL for ages 2–3 (CBCL/2–3) is a 99 item empirically validated checklist of specific problem behaviours derived from the CBCL/4–18. This was completed by a parent/ guardian at the 2-year follow-up. The CBCL is used to measure behaviour as a continuous score, in addition to apportioning factors into clinical syndrome scales including, withdrawal, anxious/depressed, somatic, social, attention, thought, aggression, and delinquency scales.24 Both the CBCL/2–3 and the CBCL/4–18 produce raw scores that can be transformed into summary T scores for total behaviour, internalising behaviour (withdrawal, anxious/depressed, and somatic), and externalising behaviours (delinquency and aggression). The recommended clinical cut-off scores (T ≥ 60) were applied to the CBCL T scores to obtain three binary variables indicative of clinically significant scores on the total, internalising, and externalising problem scales.24 In this study, scoring above this clinical cut-off is referred to as behavioural morbidity.
Predictor variables Maternal and paternal age was considered as both continuous and categorical variables. For the categorical models, maternal and paternal age at birth of the study child was classified within 5-year age categories as per those widely used in population fertility data:1 $24 000 AUD per annum/ ≤$24 000 AUD per annum; threshold representing the poverty line at the time26), and gender of offspring.
Statistical analyses Logistic regression models with generalised estimating equations were used to generate odds ratios which estimate the effect of maternal and paternal age on clinically significant T scores in offspring from age 2 to 17 years. Odds ratios when parental age was included as a categorical measure were generated relative to a reference category of parents aged 25– 29 years. An exchangeable correlation matrix was selected for the models. These models were used to account for repeated observations of the same individuals over time. The first model examined maternal and paternal age and offspring behavioural outcome while accounting for age of other parent and the confounding variables listed above. We selected confounding variables known to be associated with child behaviour outcomes, and all variables were entered into the final adjusted model. We then ran the model selecting only for first-time mothers. Ostensibly, the psychosocial profile of a mother of advanced age having her first baby is quite different to that of a mother of advanced age having another baby, which may impact upon the relationship of interest. The
44
J. E. Tearne et al. gender all accounted for a significant proportion of the variance in all three models (P = 0.001) such that offspring who were male, had mothers who did not finish high school, had mothers who smoked during pregnancy, or were born to families living below the poverty line were at significantly increased risk of clinical levels of behaviour problems. Maternal ethnicity, parity, and alcohol use during pregnancy were not significantly associated with behaviour outcomes. Correlations between raw CBCL scores at each follow up are presented in Table 2.
results of an unadjusted model examining only the relationship between maternal and paternal age and offspring outcome with adjustment for age of other parent are presented in Table S2. IBM SPSS Statistics 19.0 was used for the analyses.
Results Frequency data for each of the variables by mothers’ age are presented in Table 1. Behavioural problems (CBCL total/internalising/externalising T-score ≥ 60; population averaged estimates of T-scores observed over each follow-up), maternal education, maternal ethnicity, smoking in pregnancy, and total family income were all significantly related to both maternal and paternal age. Maternal and paternal age were strongly correlated with each other, r = 0.61. Preliminary analyses indicated that there was not a significant interaction between mothers’ and father’s age and behavioural outcomes, and this interaction was not retained in further analyses. Maternal education, total family income, maternal smoking in pregnancy, and
Linear models When used as a continuous variable, there was a significant linear relationship between maternal age and total, internalising, and externalising behaviour morbidity in offspring (Table 3). Increasing maternal age was associated with a decreasing risk for problem behaviours in offspring. No significant relationships between paternal age and behaviour outcomes were noted.
Table 1. Characteristics of participants by mothers’ age Maternal age (years)
Pa Behavioural problems (% with T score ≥ 60) Total Internalising Externalising Maternal education at pregnancy 12 or more years education 24 000 per annum ≤24 000 per annum Gender of child Male Female a
