SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

Aim  The aim of this study was to examine the association between maternal alcohol use disorder and intellectual disability in children.

Method  All mothers with an International Classification of Diseases (ICD) 9 and/or 10 alcohol-related diagnosis, a proxy for alcohol use disorder, recorded on the Western Australian health, mental health, and drug and alcohol data sets were identified through the Western Australian Data Linkage Unit (n=5614 non-Aboriginal; n=2912 Aboriginal). A comparison cohort of mothers without an alcohol-related diagnosis was frequency matched on maternal age within maternal Aboriginal status and year of birth of their children. Linkage with the Western Australian Midwives Notification System (1983–2001) identified all births to these mothers (n=10 664 and 7907 respectively). Linkage to the Western Australian Intellectual Disability Database and Register of Developmental Anomalies identified cases of intellectual disability with no identified genetic origin (intellectual disability) (n=1487) and fetal alcohol syndrome (n=66). Odds ratios (ORs) and 95% confidence intervals (CIs) for intellectual disability were calculated using logistic regression incorporating generalized estimating equations and used to estimate population-attributable fractions.

Results  At least 3.8% (95% CI 2.84–4.89%) of cases of intellectual disability could be avoided by preventing maternal alcohol use disorder: 1.3% (95% CI 0.81–1.86%) in non-Aboriginal and 15.6% (95% CI 10.85–20.94%) in Aboriginal children. We observed a three-fold increase in the adjusted odds of intellectual disability in children of mothers with an alcohol-related diagnosis recorded during pregnancy (non-Aboriginal OR 2.89, 95% CI 1.62–5.18; Aboriginal OR 3.12, 95% CI 2.13–4.56), with a net excess proportion of 3.7% and 5.5% respectively. One-third (32%) of children diagnosed with fetal alcohol syndrome had intellectual disability.

Interpretation  Maternal alcohol use disorder is the leading known risk factor for intellectual disability with no identified genetic origin.


Abbreviations
ARND

Alcohol-related neurodevelopmental disorder

FAS

Fetal alcohol syndrome

FASD

Fetal alcohol spectrum disorder

GEE

Generalised estimating equation

ICD

International Classification of Disease

IDEA

Intellectual Disability Exploring Answers

What this paper adds

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information
  •  Maternal alcohol use disorder accounts for at least 3.8% of all cases of intellectual disability.
  •  Children of mothers with an alcohol-related diagnosis have a three-fold increased risk of intellectual disability.
  •  The proportion of avoidable intellectual disability is at least 1.3% in non-Aboriginal children and 15.6% in Aboriginal children.

‘Fetal Alcohol Syndrome (FAS) is the leading non-genetic cause of intellectual disability.’1 This statement rapidly became the catchcry following Abel and Sokol’s publications in the late 1980s.1,2 The authors estimated that the worldwide incidence of FAS was 1.9 cases per 1000 live births, and they postulated that FAS-related intellectual disability accounted for 11% of institutional costs for intellectually disabled residents in the USA.2 Yet this statement has never been ratified using population-based research, and the contribution of maternal alcohol use disorder to the burden of intellectual disability remains unknown.

The assumption appears to be that all children with FAS have intellectual disability. However, the current evidence does not support this. A wide range of IQs has been observed in children with FAS and broader fetal alcohol spectrum disorders (FASDs).3–8 The mean IQ of individuals with FAS is reported to be around 68–70, with one-third classified as intellectually disabled (IQ<70).7,8 Intellectual disability is only one of a wide range of fetal effects classified as FASDs. Based on evidence from school-based studies, the highest documented prevalence of FAS and partial FAS ranges from 2% to 5%.9 The prevalence of alcohol-related neurodevelopmental disorder (ARND)10 is more difficult to ascertain since the characteristic FAS facial features are absent. Estimates of the prevalence of ARND are considerably higher than for FAS or partial FAS.9 Although the mean IQ of children with FASD in these studies was lower than in children with ARND (mean IQ 85 and 109 receptively), the prevalence of intellectual disability and the attributable fractions were not documented.9 These knowledge gaps have critical implications for health professionals, who are unable to advise what percentage of children prenatally exposed to heavy levels of alcohol will have an intellectual disability.

Ascertaining population-based estimates of the proportion of intellectual disability that can be attributed to maternal alcohol use disorder occurring during pregnancy is problematic. In the absence of the characteristic FAS dysmorphology, prenatal alcohol exposure may not be identified as a contributing cause of intellectual disability in children. This is likely to be the case in Australia, where pregnant mothers are not routinely asked about alcohol consumption during pregnancy,11 health professionals lack knowledge about FAS,11,12 and women who consume high levels of alcohol are difficult to retain in longitudinal studies.13

The present study overcomes these limitations by using linked, routinely collected population-based data to examine the association between the presence of an alcohol-related diagnosis in maternal health data sets, a proxy for maternal alcohol use disorder, and the presence of an intellectual disability with no identified genetic origin in the children of the identified mothers.

Method

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

Cohort selection

The cohort selection has been described previously.14 All women with a birth recorded on the Western Australian Midwives Notification System between 1983 and 2007 were eligible for selection into the study (n=253 714 women; n=242 956 [95.8%] non-Aboriginal women, n=10 758 [4.2%] Aboriginal women). The Midwives Notification System is a statutory notification system that records all births in Western Australia occurring at 20 weeks’ gestation or later or of infants of 400g or more birthweight. The exposed cohort comprises women in whom an alcohol-related diagnosis (International Classification of Diseases, revision 9 and/ or 10 [ICD-9 and/or ICD-10]) was recorded through routinely collected data on the Western Australian administrative Hospital Morbidity Data System (hospital in-patients), Mental Health Outpatients, and the Perth-based Drug and Alcohol Office data sets, and all their offspring whose births were recorded on the Midwives Notification System. The presence of an alcohol-related diagnosis on any one of these data sets was used as a proxy for a maternal alcohol use disorder, indicating heavy alcohol use. The diagnoses included acute alcohol intoxication, alcohol dependence, alcohol-related psychiatric conditions, alcohol-related disease, and fetal harm in the offspring (recorded on the maternal record; (Table SI, online supplementary information).14 The majority of the exposed mothers had received at least one diagnosis of acute intoxication (unpublished data).

The exposed cohort was frequency matched on maternal age within maternal race (non-Aboriginal: ≅3 unexposed to 1 exposed; Aboriginal: ≅2 unexposed to 1 exposed) and their child’s year of birth to randomly selected mothers on the Midwives Notification System who had never had an alcohol-related diagnosis recorded on the administrative data sets. The comparison cohort comprised these mothers and all their children whose births were recorded on the Midwives Notification System. Records for the exposed and comparison cohorts were linked by the Western Australian Data Linkage Branch, using probabilistic matching.15

The children’s data were linked with the Intellectual Disability Exploring Answers (IDEA) database, a register of all cases of intellectual disability in Western Australia and one of the few databases of its kind in the world. Two sources provide information about cases of intellectual disability: the Disability Services Commission and the Western Australian Department of Education.16,17 As children with mild intellectual disability may not be identified until they enter school, this study was restricted to births from 1983 to 2001 (inclusive) so that all children had reached 6 years of age and were attending the first year of primary school. The Disability Services Commission defines intellectual disability as an IQ of less than 70, with the level of severity classified according to the DSM-IV recommendations18 as ‘mild’ (IQ 50–55 to 69), ‘moderate’ (IQ 35–40 to 40–54), or ‘severe–profound’ (IQ <35 or 40).19 The Western Australian Department of Education defines intellectual disability as having intellectual functioning that is 2SD or more below the mean on an approved measure of cognitive functioning, with similar levels for severity as the Disability Services Commission scale but with mild and moderate collapsed into one category. Therefore, two levels of severity are used in this study, mild–moderate and severe.

Birth data were also linked to the Western Australian Register of Developmental Anomalies20 (previously called the Western Australian Birth Defects Register), which collects information for the Western Australian population on birth defects diagnosed in stillbirths, terminations of pregnancy, and live births up to 6 years of age, using multiple sources of ascertainment.21,22

The IDEA database has information on the cause of intellectual disability, where available, categorized into a broader group of biomedical causes or otherwise, based loosely on terminology used by Yeargin-Allsopp et al.,23 where biomedical diagnoses include genetic conditions (such as Down syndrome). Children with a diagnosis of FAS were identified through linkage with the Western Australian Register of Developmental Anomalies.24

Chromosomal causes of intellectual disability (British Paediatric Association Codes 75800–75899) were identified through the Western Australian Register of Developmental Anomalies and the IDEA database and excluded from analysis.

Statistical analysis

The proportion of cases of intellectual disability, including by severity of intellectual disability, was calculated separately for non-Aboriginal and Aboriginal children. Results are presented per 1000 live births and the net excess proportion per 1000 live births is calculated for the exposed cohort. The proportion of intellectual disability in children with a diagnosis of FAS is presented per 1000 children with FAS. Where the number of cases within a stratum is fewer than five, results are not presented to prevent possible identification of individuals.

Classification of the timing of maternal alcohol-related diagnosis

The pregnancy period was estimated by subtracting gestational age16 at birth from date of birth to give the date of conception. Mothers with any alcohol-related diagnosis recorded during pregnancy, which may also include an alcohol-related diagnosis before and/or after pregnancy, were classified as (i) ‘during pregnancy’.14 For women without a diagnosis during pregnancy, the maternal alcohol-related diagnosis was coded into the timing of recording of the alcohol-related diagnosis in relation to pregnancy using a hierarchical coding (Table SII, online supporting information) so that diagnoses recorded within 1 year of pregnancy were coded as (ii) up to 1 year before pregnancy and may include exposure more than 1 year before pregnancy or any exposure after pregnancy, or (iii) up to 1 year after pregnancy and may include exposure more than 1 year before or after pregnancy. Mothers with a diagnosis only recorded more than 1 year either side of pregnancy were grouped as (iv) more than 1 year before pregnancy and may include exposure more than 1 year after pregnancy or (v) more than 1 year after pregnancy.14 Maternal alcohol-related diagnoses were also coded into a binary variable (yes/no) and a separate group in which the mother had an alcohol-related diagnosis both pre and post pregnancy, but not during pregnancy.

The odds of intellectual disability were calculated for the exposed offspring compared with offspring in the comparison cohort, stratified by Aboriginal status, using generalized estimating equations (GEEs), which take into account the correlation between siblings.25 We re-ran the analysis after removing cases of FAS to examine if the risk of intellectual disability was attenuated. GEE analyses were conducted using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). Results are presented as odds ratios with 95% confidence intervals (CIs).

All GEE analyses were adjusted for the factors used in frequency matching (maternal age and year of birth of the mother’s children). Other potential confounders included maternal demographic characteristics – marital status, parity, maternal illicit drug use (any ICD9/10 code for illicit drugs on the data sets), health region within Western Australia at time of childbirth (Perth metropolitan [metro], rural/remote Western Australia, based on postcode), socio-economic status and index of education and occupation at the time of childbirth based on Australian Census data collected at the census district level (approximately 200 households)26– and stratified into quintiles and mental health diagnoses (ICD9/10 codes on the data sets) grouped into three separate variables: schizophrenia, depression, and all other diagnoses. Each of these variables was entered into the base model and those that altered the odds ratio by 20% or more were included as confounders. The only variable to alter the odds ratio by 20% or more was ‘other’ mental health diagnoses for non-Aboriginal but not for Aboriginal analyses. Hence, non-Aboriginal odds ratios were also adjusted for this variable and Aboriginal analyses were adjusted only for matching variables.

Population-attributable fractions and 95% CIs were calculated for any alcohol-related diagnosis and exposure during pregnancy, by Aboriginal status,27 using whole Western Australia population numbers of all intellectual disability for the comparison numerators and denominators, obtained from the IDEA database. Numbers of mothers and births were obtained from the Midwives Notification System. The population-attributable fraction is the percentage of intellectual disability that could be prevented by eliminating maternal alcohol use disorders during pregnancy.28

Results

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

There were 64 842 children in the study; 41 320 (63.7%) of the children were non-Aboriginal, of whom 10 576 (25.6%) had a mother with an alcohol-related diagnosis; 23 522 (36.3%) were Aboriginal, of whom 7 760 (33.0%) had a mother with an alcohol-related diagnosis (Table I). Exposed mothers were less likely than comparison mothers to be married, and more likely to be of higher parity, have a mental health diagnosis, or have a diagnosis of illicit drug use. The majority of non-Aboriginal mothers in both the exposed and comparison cohorts lived in Perth (72.2%), with the majority of Aboriginal mothers living in rural/remote Western Australia (69.6%) at the time of birth of their child (Table I).

Table I. Maternal demographic characteristics by Aboriginal status, for each live birth between 1983 and 2001
 Non-Aboriginal births (n=41 320)Aboriginal births (n=23 522)
Exposed (n=10 576), n (%)Comparison (n=30 744), n (%)Exposed (n=7760), n (%)Comparison (n=15 762), n (%)
  1. WA, Western Australia.

Maternal age (y)
 <201377 (13.0)4024 (13.1)2297 (29.6)4238 (26.9))
 20–243284 (31.1)9514 (30.9)2681 (34.5)5605 (35.5)
 25–293184 (30.1)9199 (29.9)1686 (21.7)3624 (23.0)
 30–341895 (17.9)5542 (18.0)806 (10.4)1667 (10.6)
 35–39707 (6.7)2111 (6.9)266 (3.4)559 (3.5)
 40+129 (1.2)354 (1.2)24 (0.3)69 (0.4)
Marital status
 Married7757 (73.5)26 286 (85.5)4377 (56.6)10 065 (64.0)
 Never married2395 (22.7)4179 (13.6)3212 (41.5)5424 (34.5)
 Divorced/widowed401 (3.8)269 (0.9)144 (1.9)229 (1.5)
Parity
 04168 (39.4)15 087 (49.1)1989 (25.6)4646 (29.5)
 13214 (30.4)9618 (31.3)1758 (22.7)3797 (24.1)
 21863 (17.6)4115 (13.4)1527 (19.7)2949 (18.7)
 3+1331 (12.6)1924 (6.3)2486 (32.0)4370 (27.7)
 Illicit drug use5343 (38.3)608 (1.5)2226 (22.4)979 (4.7)
Mental health
 Schizophrenia473 (4.5)88 (0.3)309 (4.0)119 (0.8)
 Depression3223 (30.5)948 (3.1)974 (12.6)749 (4.7)
 All other diagnoses5171 (48.9)2239 (7.3)2081 (26.8)1624 (10.3)
Health region of WA
 Perth metropolitan7625 (72.1)21 809 (70.9)2210 (28.5)4887 (30.9)
 Rural/remote WA2895 (27.4)8675 (28.2)5535 (71.3)10 824 (68.7)
 Outside WA56 (0.5)260 (0.8)15 (0.2)51 (0.3)

Intellectual disability was diagnosed in 1660 children. A genetic/chromosomal cause was identified for 169 children in the IDEA database, and for a further four children a chromosomal cause was identified from linkage to the Western Australian Register of Developmental Anomalies (British Paediatric Association Codes 75800–75899). This equated to 10.4% of all intellectual disability in this cohort having a known genetic/chromosomal cause. These 173 cases were excluded from this study, leaving 1487 cases of intellectual disability with no identified genetic origin, referred to as intellectual disability in this paper (Table II).

Table II. Proportion of intellectual disability per 1000 live births between 1983 and 2001, by maternal alcohol exposure and Aboriginal status
Intellectual disability severityNon-aboriginal (n=41 320, 63.7%)Aboriginal (n= 23 522, 36.3%)
Any alcohol-related diagnosis (n=10 576)Comparison (n=30 744)Net excessbAny alcohol-related diagnosis (n=7760)Comparison (n=15 762)Net excessb
n Proportion (%)a (95% CI) n Proportion (%)a (95% CI) n Proportion (%)a (95% CI) n Proportion (%)a (95% CI)
  1. aProportion per 1000 live births. bNet excess: proportion in the exposed cohort minus the proportion in the comparison cohort. cFive children had an unspecified level of intellectual disability; missing data 0.3%. CI, confidence interval.

All intellectual disabilitiesc26525.1 (22.2–28.2)41113.4 (12.1–14.7)11.7 (8.6–15.1)35846.1 (41.7–51.0)45328.7 (26.2–31.5)17.4 (12.2–22.9)
Mild–moderate24723.4 (20.6–26.4)38312.5 (11.3–13.8)10.9 (7.9–14.2)34644.6 (40.2–49.4)42026.6 (24.2–29.3)18.0 (12.8–23.3)
Severe171.6 (1.0–2.6)260.8 (0.6–1.2)0.8 (0.0–1.8)111.4 (0.8–2.5)332.1 (1.5–2.9)–0.7 (–1.7 to 0.6)

The proportion of intellectual disability in all exposed non-Aboriginal children was 25.1 per 1000 live births (95% CI 22.2–28.2) and 13.4 per 1000 (95% CI 12.1–14.7) for the comparison children, with a net excess proportion of 11.7 per 1000 live births (Table II). For Aboriginal children, the proportion was 46.1 per 1000 live births (95% CI 41.7–51.0) in the exposed cohort and 28.7 per 1000 live births (95% CI 26.2–31.5) in the comparison cohort, giving a net excess proportion of 17.4 per 1000 live births. In the majority of cases, the intellectual disability was classified as mild–moderate. The net excess proportion of severe intellectual disability in exposed children was 0.8 per 1000 live births (95% CI 0.0–1.8) for non-Aboriginal children and −0.7 per 1000 live births (95% CI −1.7 to 0.6) for Aboriginal children, with the proportion of severe intellectual disability in Aboriginal comparison children being more than twice that of the non-Aboriginal comparison children (2.1 vs 0.8; Table II).

There were 66 children (0.4% of exposed) with a diagnosis of FAS recorded on the Western Australia Register of Developmental Anomalies; eight (12%) were non-Aboriginal and 58 (88%) were Aboriginal (results not tabled). Approximately one-third (31.8%) of children with FAS had received a diagnosis of intellectual disability, giving a proportion of 318 cases of intellectual disability per 1000 children with FAS, of whom 90% had mild–moderate intellectual disability.

When a maternal alcohol-related diagnosis was recorded during pregnancy, the overall proportion of intellectual disability in non-Aboriginal children was 50.8 per 1000 live births (95% CI 29.9–84.9) with a net excess proportion of 37.4 per 1000 live births (Table III). A smaller increase in excess proportions was evident when a maternal diagnosis was recorded either pre or post pregnancy but not during pregnancy. The proportion of intellectual disability in the offspring of mothers with alcohol diagnoses recorded both pre and post pregnancy, but not during pregnancy, was 37.2 per 1000 live births (95% CI 25.5–54.0) with a net excess of 23.8 per 1000 live births.

Table III. Proportion of intellectual disability per 1000 live births between 1983 and 2001, by timing of maternal alcohol-related diagnosis and Aboriginal status
Timing of exposureNon-aboriginal (= 10 576)Aboriginal (= 7760)
na (%) nb Proportion (%)c (95% CI)Net excessd na (%) nb Proportion (%)c (95% CI)Net excessd
  1. aNumber of non-Aboriginal/Aboriginal children in group. bTotal number of children with intellectual disability in each group. cProportion per 1000 live births. dNet excess: proportion in exposed cohort minus the proportion in the comparison cohort for intellectual disability in Table II. CI, confidence interval.

During pregnancy256 (2.4)1350.8 (29.9–84.9)37.4 (16.5–71.6)443 (5.7)3783.7 (61.3–113.3)54.8 (32.3–84.4)
≤1y pre pregnancy518 (4.9)1019.3 (10.5–35.2)5.9 (3.0–21.8)358 (4.6)2467.0 (45.5–97.8)38.3 (16.5–69.2)
≤1y post pregnancy280 (2.6)621.4 (9.9–46.0)8.0 (3.6–32.6)310 (4.0)1858.1 (37.0–89.9)29.4 (8.1–61.3)
>1y pre pregnancy1770 (16.7)5329.9 (23.0–39.0)16.6 (9.5–25.7)779 (10.0)4152.6 (39.0–70.6)23.9 (10.0–42.1)
>1y post pregnancy7752 (73.3)18323.6 (20.5–27.2)10.2 (6.8–14.1)5843 (75.3)23840.7 (36.0–46.1)12.0 (6.5–17.9)
Pre and post pregnancy (not during pregnancy)698 (6.6)2637.2 (25.5–54.0)23.8 (12.1–40.7)930 (12.0)5963.4 (49.5–81.0)34.7 (20.5–52.4)

For Aboriginal children, the overall proportion of intellectual disability was 83.7 per 1000 live births (95% CI 61.3–113.3) for an alcohol-related diagnosis recorded during pregnancy, giving a net excess proportion of 54.8 per 1000 live births (Table III). The excess proportions for each of the other categories without an alcohol-related diagnosis recorded during pregnancy were higher than for non-Aboriginal offspring, ranging from 12.0 per 1000 live births when an alcohol-related diagnosis was recorded more than 1 year post pregnancy to 38.3 per 1000 live births when an alcohol-related diagnosis was recorded within 1 year pre pregnancy. The proportion of intellectual disability in the offspring of mothers with alcohol diagnoses recorded both pre and post pregnancy, but not during pregnancy, was 63.4 per 1000 live births (95% CI 49.5–81.0), with a net excess of 34.7 per 1000 live births (Table III).

Non-Aboriginal children of mothers with any alcohol-related diagnosis had increased odds of intellectual disability compared with comparison children (adjusted OR [aOR] 1.44; 95% CI 1.18–1.75; Table IV). Increased odds were found when an alcohol-related diagnosis was recorded during pregnancy (aOR 2.89; 95% CI 1.62–5.18) and when diagnoses were recorded more than 1 year pre pregnancy (aOR1.87; 95% CI 1.36–2.57). When alcohol diagnoses were recorded both pre and post pregnancy, but not during pregnancy, the odds ratios doubled (aOR 1.95; 95% CI 1.25–3.05). These results remained valid after removing children with FAS from the analyses (Table IV).

Table IV. Odds ratios for intellectual disability by maternal alcohol-related diagnosis and Aboriginal status (for live births between 1983 and 2001)
 Non-aboriginalNon-aboriginal, no FASc, ORb (95% CI)Aboriginal, ORa (95% CI)Aboriginal, no FASc, ORa (95% CI)
ORa (95% CI)ORb (95% CI)
  1. aAdjusted for variables used in frequency matching, maternal age, and year of birth. bNon-Aboriginal analyses adjusted for frequency matching variables maternal age and year of birth and maternal mental health ‘other’. cCases of FAS removed from analysis. FAS, fetal alcohol syndrome; OR, oddsratio; CI, confidence interval.

Any alcohol diagnosis
 No1.001.001.001.00 
 Yes1.81 (1.53–2.14)1.44 (1.18–1.75)1.43 (1.17–1.74)1.66 (1.42–1.96)1.58 (1.34–1.86)
Timing of alcohol diagnosis
 No alcohol diagnosis1.001.001.001.001.00
 During pregnancy3.52 (1.96–6.34)2.89 (1.62–5.18)2.71 (1.48–4.98)3.12 (2.13–4.56)2.71 (1.81–4.07)
 ≤1y pre pregnancy1.15 (0.60–2.20)0.90 (0.47–1.73)0.90 (0.47–1.73)2.30 (1.49–3.55)2.02 (1.28–3.19)
 ≤1y post pregnancy1.52 (0.70–3.30)1.17 (0.54–2.53)1.17 (0.54–2.52)1.95 (1.19–3.20)1.75 (1.04–2.94)
 >1y pre pregnancy2.27 (1.67–3.09)1.87 (1.36–2.57)1.86 (1.35–2.56)2.23 (1.56–3.19)2.20 (1.54–3.16)
 >1y post pregnancy1.71 (1.41–2.07)1.33 (1.07–1.66)1.32 (1.06–1.65)1.44 (1.20–1.73)1.39 (1.16–1.67)
 Pre and post pregnancy (not during pregnancy)2.71 (1.76–4.16)1.95 (1.25–3.05)1.95 (1.25–3.05)2.22 (1.62–3.02)1.97 (1.42–2.75)

The relationship between a maternal alcohol-related diagnosis and intellectual disability was more striking in Aboriginal children than in non-Aboriginal children (Table IV). The aOR for intellectual disability in Aboriginal children of mothers with any alcohol-related diagnosis was 1.66 (95% CI 1.42–1.96), and increased odds ratios were observed for each of the six time periods. The odds ratio was increased three-fold when a diagnosis was recorded during pregnancy (aOR 3.12; 95% CI 2.13–4.56), and doubled when an alcohol-related diagnosis was recorded not during pregnancy but up to 1 year before pregnancy or more than 1 year prior to pregnancy, or when alcohol diagnoses were recorded both pre and post pregnancy. As with the non-Aboriginal analyses, the results remained valid following removal of children with FAS (Table IV).

The population-attributable fraction of intellectual disability when a maternal diagnosis occurred during pregnancy was 0.14% (95% CI 0.02–0.29%) for non-Aboriginal children and 2.8% (95% CI 1.09–4.68%) for Aboriginal children (results not shown). The population-attributable fraction for any alcohol-related diagnosis was 1.3% (95% CI 0.81–1.86%) for non-Aboriginal and 15.6% (95% CI 10.85–20.94%) for Aboriginal children. When combined, the population-attributable fraction equates to 3.8% (95% CI 2.84–4.89) of all cases of intellectual disability in Western Australia.

Discussion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

These results provide the first population-based evidence that a maternal alcohol-related diagnosis, which can be considered a proxy for maternal alcohol use disorder, is the leading known cause of intellectual disability with no identified genetic origin. In total, 3.8% (95% CI 2.84–4.89%) of all cases of intellectual disability in Western Australia were associated with maternal alcohol use disorder. Furthermore, 1.3% (95% CI 0.81–1.86%) of intellectual disability in non-Aboriginal children and 15.6% (95% CI 10.85–20.94) in Aboriginal children of mothers with an alcohol-related diagnosis could potentially be prevented by eliminating heavy alcohol use by these mothers during pregnancy.

The net excess proportion of intellectual disability provides an estimate of the proportion of intellectual disability occurring in the children of mothers with an alcohol-related diagnosis. Among the pregnancies during which an alcohol-related diagnosis was recorded in the mother, 3.7% (95% CI 1.65–7.16%) of the children of non-Aboriginal mothers and 5.5% (95% CI 3.23–8.44%) of children of Aboriginal mothers had an intellectual disability. This is the first study to provide health professionals with evidence on which to base their clinical advice on the risk of intellectual disability in the children of women drinking heavily in pregnancy.

Calculation of the population-attributable fraction requires a large population-based study, and few studies investigating risk factors for intellectual disability have this capacity. We are not aware of any other behavioural risk factor for intellectual disability that has an attributable fraction greater than the 3.8% reported for the children of mothers with an alcohol-related diagnosis, and in particular the attributable fraction of 15.6% for Aboriginal children. We recognize that these are conservative estimates. Cases of intellectual disability on the IDEA database can be ascertained up to 18 years of age, with 80% identified by 10 years of age.17 Twenty per cent of the cohort in this study was between 6 and 9 years of age at the end of follow-up in 2007, so some cases of intellectual disability would not have been identified. Also, as these results relate to the children of mothers with an alcohol use disorder that has been identified in the health setting, they represent the extreme end of the exposure continuum.

As the risk of intellectual disability was the same for exposed non-Aboriginal and Aboriginal cohorts, the population-attributable fraction will be influenced by the prevalence of the exposure in these communities. The reported prevalence of alcohol use disorders in non-Aboriginal Australian women is 3%29 to 5.0%,30 higher than the 2.3% obtained in this study (unpublished data), which would generate a higher population-attributable fraction for intellectual disability than the estimate of 1.3% obtained in this study. We are unaware of any estimates of the prevalence of alcohol use disorders for Aboriginal mothers. However, it is possible that underascertainment of Aboriginal mothers with an alcohol-use disorder also occurred with this study design. Alcohol-related problems are more common in rural/remote regions than in metropolitan regions in Australia;31 the majority of Aboriginal mothers in this study lived in rural or remote regions, and we were unable to access service data for Aboriginal-specific health services and rural drug and alcohol services for this study. It is also possible that Aboriginal mothers are more open about their drinking, making it easier for health professionals to identify, or that health professionals are more likely to record an alcohol diagnosis for Aboriginal than for non-Aboriginal mothers; however, we are unable to examine these issues in this study.

A measure of the under-recognition of alcohol use disorders during pregnancy is the lower attributable fraction obtained for intellectual disability when a maternal alcohol-related diagnosis was recorded during pregnancy. The estimates of 0.14% for non-Aboriginal children and 2.8% for Aboriginal children when a maternal alcohol-related diagnosis was recorded during pregnancy are considerably less than the 1.3% and 15.6% respectively, obtained for an alcohol-related diagnosis recorded at any time of a woman’s life. Health professionals have a key role to play in identifying heavy alcohol consumption in women of childbearing age and pregnant women and implementing brief interventions to reduce the risk of alcohol-exposed pregnancies.32 These are essential for preventing alcohol-related intellectual disability.

The risk of intellectual disability in children of mothers with an alcohol-related diagnosis transcends race. We acknowledge that for some Aboriginal children, particularly those in remote regions, assessments of intellectual disability may be culturally inappropriate, putting these children at a disadvantage. However, psychologists in Western Australia are generally aware of the need to use ‘culture-free’ tools for Aboriginal children.33 In both non-Aboriginal and Aboriginal children, the odds of intellectual disability were increased three-fold when an alcohol-related diagnosis was recorded for the mother during pregnancy, and this increase remained after adjusting for potential confounding factors and following removal of cases of FAS. The prevalence of FAS in Western Australia (0.5/1000 births)24 is lower than many published estimates (2–7/1000).9 There is a lack of recognition of FAS in Western Australia11 and limited diagnostic capacity so it is probable that other exposed children with intellectual disability in this cohort have undiagnosed FAS. However, it is likely that some of the exposed children with intellectual disability will lack the classic FAS facial features and therefore fit the diagnostic criteria for an ARND. So, given that FAS is less common than maternal alcohol use disorder,9,29,30 not all children with FAS have intellectual disability, and that many children with alcohol-related intellectual disability may be more appropriately classified as having an ARND, it is more appropriate to state that maternal alcohol use disorder is the leading known non-genetic cause of intellectual disability, rather than FAS, as suggested by Abel and Sokol 30 years ago.1,2

Linkage of routinely collected population-based data is a valid means of identifying cases admitted to hospital for a health-related condition34 and minimizes loss to follow-up of high-risk mothers and their children,13,35 thereby overcoming many of the difficulties inherent in epidemiological studies investigating maternal alcohol consumption during pregnancy and intellectual disability.16,19 However, the majority of alcohol diagnoses were for acute intoxication and we could not examine the exact timing of exposure and whether there were extended periods of sobriety which could have reduced risk to the fetus. Overall, the proportion of intellectual disability in the non-Aboriginal comparison cohort (13.4/1000) was similar to the published Western Australian prevalence,17 and the proportion of intellectual disability in children with FAS (318/1000 children with FAS) was consistent with published estimates.7 This consistency supports the validity of the data.

Conclusion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

These results provide the first population-based evidence to support and clarify the conclusions made by Abel and Sokol in 1986,1 indicating that maternal alcohol use disorder should be classified as the leading known cause of intellectual disability with no identified genetic origin. Prevention of maternal alcohol use disorder during pregnancy could, conservatively, avoid 3.8% of all cases of intellectual disability, and in Aboriginal children 15.6% of intellectual disability could be avoided. Early identification of heavy drinking by women of childbearing age and pregnant women and brief interventions are fundamental prevention strategies.

References

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information
  • 1
    Abel EL, Sokol RJ. Fetal alcohol syndrome is now leading cause of mental retardation. Lancet1986; 2: 1222.
  • 2
    Abel EL, Sokol RJ. Incidence of fetal alcohol syndrome and economic impact of FAS-related anomalies. Drug Alcohol Depend1987; 19: 5170.
  • 3
    Aragon A, Coriale G, Fiorentino D, et al.Neuropsychological characteristics of Italian children with fetal alcohol spectrum disorders. Alcohol Clin Exp Res2008; 32: 190919.
  • 4
    Ervalahti N, Korkman M, Fagerlund A, Autti-Ramo I, Loimu L, Hoyme HE. Relationship between dysmorphic features and general cognitive function in children with fetal alcohol spectrum disorders. Am J Med Genet A2007; 143A: 291623.
  • 5
    Jacobson SW, Jacobson JL, Sokol RJ, Chiodo LM, Corobana R. Maternal age, alcohol abuse history, and quality of parenting as moderators of the effects of prenatal alcohol exposure on 7.5-year intellectual function. Alcohol Clin Exp Res2004; 28: 173245.
  • 6
    Kodituwakku P, Coriale G, Fiorentino D, et al.Neurobehavioral characteristics of children with fetal alcohol spectrum disorders in communities from Italy: preliminary results. Alcohol Clin Exp Res2006; 30: 155161.
  • 7
    Steinhausen HC, Spohr HL. Long-term outcome of children with fetal alcohol syndrome: psychopathology, behavior, and intelligence. Alcohol Clin Exp Res1998; 22: 33438.
  • 8
    Mattson S, Riley E, Gramling L, Delis D, Jones K. Heavy prenatal alcohol exposure with or without physical features of fetal alcohol syndrome leads to IQ deficits. J Pediatr1997; 131: 71821.
  • 9
    May PA, Gossage JP, Kalberg WO, et al.Prevalence and epidemiologic characteristics of FASD from various research methods with an emphasis on recent in-school studies. Dev Disabil Res Rev2009; 15: 17692.
  • 10
    Stratton K, Howe C, Bataglia F. Fetal Alcohol Syndrome. Diagnosis, Epidemiology, Prevention, and Treatment. Washington, DC: National Academies Press, 1996.
  • 11
    Payne J, France K, Henley N, et al.Changes in health professionals’ knowledge, attitudes and practice following provision of educational resources about prevention of prenatal alcohol exposure and fetal alcohol spectrum disorder. Paediatr Perinat Epidemiol2011; 25: 31627.
  • 12
    Abel EL, Kruger M. What do physicians know and say about fetal alcohol syndrome: a survey of obstetricians, paediatricians, and family medicine physicians. Alcohol Clin Exp Res1998; 22: 195154.
  • 13
    O’Callaghan FV, O’Callaghan M, Najman JM, Williams GM, Bor W. Prenatal alcohol exposure and attention, learning and intellectual ability at 14 years: a prospective longitudinal study. Early Hum Dev2007; 83: 11523.
  • 14
    O’Leary CM, Watson L, D’Antoine H, Stanley F, Bower C. Heavy maternal alcohol consumption and cerebral palsy in the offspring. Dev Med Child Neurol2012; 54: 22430.
  • 15
    Holman CD, Bass AJ, Rouse IL, Hobbs MS. Population-based linkage of health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health1999; 23: 45359.
  • 16
    Petterson B, Leonard H, Bourke J, et al.IDEA (Intellectual Disability Exploring Answers): a population-based database for intellectual disability in Western Australia. Ann Hum Biol2005; 32: 23743.
  • 17
    Leonard H, Peterson P, Bourke J, Morgan V, Glasson E, Bower C. Inaugural Report of the idEA Database – Intellectual Disability in Western Australia. Perth: Telethon Institute for Child Health Research, 2004.
  • 18
    American Psychiatric Association, editor. Diagnostic and Statistical Manual of Mental Disorders DSM-IV-TR, 4th edn. Washington, DC: American Psychiatric Association, 1994.
  • 19
    Leonard H, Patterson B, Bower C, Saunders R. Prevalence of intellectual disability in Western Australia. Paediatr Perinat Epidemiol2003; 17: 5867.
  • 20
    Bower C, Rudy E, Callaghan A, Quick J, Cosgrove P. Report of the Births Defects Registry of Western Australia, 1980–2007; Report No. 15. Perth: King Edward Memorial Hospital, 2009.
  • 21
    Bower C, Ryan A, Rudy E. Ascertainment of pregnancies terminated because of birth defects: effect on completeness of adding a new source of data. Teratology2001; 63: 2325. [Erratum appears in Teratology 2001 Mar; 63: 164].
  • 22
    Bower C, Silva D, Henderson TR, Ryan A, Rudy E. Ascertainment of birth defects: the effect on completeness of adding a new source of data. J Paediatr Child Health2000; 36: 57476.
  • 23
    Yeargin-Allsopp M, Murphy CC, Decoufle P, Hollowell JG. Reported biomedical causes and associated medical conditions for mental retardation among 10-year-old children, metropolitan Atlanta, 1985 to 1987. Dev Med Child Neurol1997; 39: 14249.
  • 24
    Bower C, Rudy E, Callaghan A, Quick J, Cosgrove P, Nassar N. Report of the Birth Defects Registry of Western Australia; 1980–2008. Perth: King Edward Memorial Hospital, 2009.
  • 25
    Armitage P, Berry G, Matthews J. Statistical Methods in Medical Research, 4th edn. Oxford: Blackwell Publishing, 2002.
  • 26
    Australian Bureau of Statistics. Socio-Economic Indexes for Areas (Report No. 2039.0). Canberra: Australian Bureau of Statistics, 2001.
  • 27
    Natarajan S, Lipsitz SR, Rimm E. A simple method of determining confidence intervals for population attributable risk from complex surveys. Stat Med2007; 26: 322939.
  • 28
    Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health1998; 88: 1519.
  • 29
    Teeson M, Baillie A, Lynskey A, Manor B, Degenhardt L. Substance use, dependence and treatment seeking in the United States and Australia: a cross-national comparison. Drug Alcohol Depend2006; 81: 14955.
  • 30
    Proudfoot HH, Baillie AJAJ, Teesson MM. The structure of alcohol dependence in the community. Drug Alcohol Depend2006; 81: 2126.
  • 31
    Miller PG, Coomber K, Staiger P, Zinkiewicz L, Toumbourou JW. Review of rural and regional alcohol research in Australia. Aust J Rural Health2010; 18: 11017.
  • 32
    Floyd RL, Sobell M, Velasquez MM, et al.Preventing alcohol-exposed pregnancies: a randomized controlled trial. Am J Prev Med2007; 32: 110. [Erratum appears in Am J Prev Med 2007 Apr; 32: 360.]
  • 33
    Kearins J. Children and cultural differences. In: Dudgeon P, Garvey D, Pickett H, editors. Working with Indigenous Australians: A Handbook for Psychologists. Perth: Gunada Press, 2000: 16776.
  • 34
    Brameld KJ, Thomas MAB, Holman CDJ, Bass J, Rouse IL. Validation of linked administrative data on end-stage renal failure: application of record linkage to a ‘clinical base population’. Aust N Z J Public Health1999; 23: 46467.
  • 35
    Clarke A, Preen DB, Ng JQ, Semmens JB, Holman CD. Is Western Australia representative of other Australian states and territories in terms of key socio-demographic and health economic indicators?Aust Health Rev2010; 34: 21015.

Supporting Information

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. References
  9. Supporting Information

Table SI: International Classification of Diseases 10 and 9 codes for alcohol diagnoses

Table SII: Coding of timing of alcohol-related diagnosis in relation to pregnancy

FilenameFormatSizeDescription
dmcn12029_sm_TableS1-S2.doc48KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.