**2. Literature search and selection**

is manifest before 18 years. [19] The level of ID is generally grouped into the five levels of mild, moderate, severe, profound and unspecified by IQ score. In research it is common to stratify ID to the following three levels defined by the American Psychiatric Association [2] (Table 1).

\*Here the person has been assessed as having ID but has not been assessed adequately to determine the level.

In terms of aetiology, ID can be broadly divided into cases of known biomedical cause and those of unknown cause. The biomedical causes may be divided into genetic and non-genetic

In addition to genetic and non-genetic causes of ASD and ID, relationships with sociodemographic factors such as a mother's education, [20, 21] immigration, [17, 22] and ethnicity, [23] have also been identified. Other reported associations involve aspects of a mother's health including physical characteristics [24] physical [25, 26] and mental health [27, 28] and health

It has also been reported that milder autistic traits are present in other family members of individuals diagnosed with ASD. This phenomenon has been coined the Broad Autism

**Descriptor/level of ID IQ score**

388 Recent Advances in Autism Spectrum Disorders - Volume I

Unspecified ID < 70\*

causes. Further subdivisions are given in Figure 1.

**Figure 1.** Commonly known aetiologies of intellectual disability

behaviours. [29, 30]

**Table 1.** Levels of intellectual disability

Mild or moderate ID 35-40 up to 69 points

Severe or profound ID < 35-40 points

The papers considered for this review resulted from a search of the Medline, Web of Knowl‐ edge, Scopus and Google scholar databases. Combinations of the search terms below were chosen.


A paper was included in the review if:


**•** It described the results of a cohort, case-control, correlation or cross-sectional study of at least 15 subjects;

found that children from higher income families were more likely to have a diagnosis of ASD. Similarly, others using family income as a marker for SES, found a significant association between high family SES and ASD in the offspring. [47] Further analyses, using the dual markers of high family income and high maternal education, found a particular association between high SES and ASD without ID. [47] Using population data and deriving SES from mother's place of residence at time of the child's birth, Australian researchers also found that ASD, ASD with ID and particularly ASD without ID were associated with higher SES. [17] The overall association between high SES and ASD without ID could result from the increased empowerment of parents of high SES to pursue a diagnosis where their children have a milder variant of ASD. [49] In families of low SES, higher functioning children with autistic traits might be informally labelled by family and contemporaries as unusual, difficult or emotionally damaged. In a comparable way, lower functioning children with autistic traits might be formally or informally given a diagnosis of ID. Others have suggested that children of lower SES parents might be more likely to be diagnosed at a later age than those of higher SES and

Pre-Existing Differences in Mothers of Children with Autism Spectrum Disorder and/or Intellectual Disability: A Review

http://dx.doi.org/10.5772/54488

391

hence not be included in studies of ASD and SES with lower ages of cut-off. [49]

dose-response fashion. [39]

to that of similar parents in less affluent areas.

Further evidence of the possible social contributions to the likelihood of an ASD diagnosis was found in a large multi-based national study in the US. [39] Undiagnosed children who met the criteria for ASD had a lower SES than children who had been previously diagnosed. [39] Arealevel SES indicators derived from census data were used in another study where the research‐ ers elucidated that increasing SES and the increasing prevalence of ASD were associated in a

King et al. [41] provided evidence that an interaction of social factors was affecting the likelihood of an ASD diagnosis. They examined factors influencing the likelihood of an ASD diagnosis using data on around five million births in Californian cohorts from 1992 to 2000. They found that an interaction between high and low level SES measures influenced the likelihood of an ASD diagnosis. Medi-Cal is a program providing medical assistance to the needy in California and these researchers used family use of Medi-Cal as a binary measure of SES. Property values in the area of a mother's residence were also used as a measure of SES. These researchers reported that children whose families were enrolled with Medi-Cal births and living in wealthier neighbourhoods were two and a half times more likely to receive a diagnosis of ASD than their counterparts living in poorer areas. [41] This could indicate that for parents of limited resources, living in a higher SES neighbourhood had benefits in terms of the likelihood of their child being diagnosed with ASD. Possibly, this results from the parents' increased access to support persons such as paediatricians and child health nurses and to educational programs such as parent classes and interventions for children, compared

In contrast, a Danish study accessing linked population data, used maternal education and parental wealth as a measure of SES and found no association between SES and ASD diagnosis. [42] In neighbouring Sweden, a population-based study published in 2012, used low income, manual occupation and less education as measures of low SES. The researchers concluded that low, not high SES, was a risk factor for ASD. [43] There may be a number of reasons for the differing findings of these studies. The universal health-care and routine screenings offered in


Eighty papers were retained for the review. We stress that these papers do not represent the entire literature pool in the area. Had we chosen different search terms or used different combinations of terms in our searches, the basis for our review would possibly have been different. The three categories for our analyses were; *socio-demographic factors, immigrant status and ethnicity* and *health and physical characteristics* and we sorted papers by these categories. An additional 61 articles were used to provide background information and possible explanations of some of the reported associations.

#### **3. Socio-demographic factors**

There are a number of considerations which impinge on the effect of socio-demographic factors on the prevalence of ASD and/or ID. Firstly, in persons with ASD with ID and those with just ID, the features overlap to some degree. A child with ASD with ID, particularly in the past, may often have been diagnosed with only ID as there have been secular changes in the identification of children with ASD. Secondly, persons who could be diagnosed with ASD without ID are most often able to function independently and may remain undiagnosed in a range of scenarios. Thirdly, the process through which children are assigned a diagnosis of ASD is much more complex than for ID. Whilst elsewhere, the gold standard might be the considered judgement of an expert clinician who had seen many patients with autism [36], in Western Australia, a diagnosis of ASD in a child requires an assessment by a team comprising a paediatrician, psychologist, and speech pathologist. Waiting times for this assessment can be prolonged in the Western Australian public system [37] and even longer in the US. [38] One effect of these considerations is that in some countries a child whose parents are socioeco‐ nomically disadvantaged may oft times be diagnosed with ID when a diagnosis of ASD could be more appropriate.

#### **3.1. Socio-economic status**

Thirteen articles researched the parental socio-economic status (SES) of children with ASD and/or ID [17, 20, 21, 39-48] and all but two [42, 43] supported a different association of SES with ASD than ID. Children with ASD were more likely to be from higher SES families but children with ID were more likely to be from lower SES families.

A range of measures of high SES were consistently associated with ASD. In a large telephone interview study in the US, family wealth was used as a measure of SES. [40] The researchers found that children from higher income families were more likely to have a diagnosis of ASD. Similarly, others using family income as a marker for SES, found a significant association between high family SES and ASD in the offspring. [47] Further analyses, using the dual markers of high family income and high maternal education, found a particular association between high SES and ASD without ID. [47] Using population data and deriving SES from mother's place of residence at time of the child's birth, Australian researchers also found that ASD, ASD with ID and particularly ASD without ID were associated with higher SES. [17]

**•** It described the results of a cohort, case-control, correlation or cross-sectional study of at

**•** It compared a characteristic of parents or mothers of children with ASD and/or ID with

**•** It assessed characteristics that were pre-existing and not likely to be a result of caring for a

**•** It used methods of ascertainment and measurement of the characteristic(s) of interest that

Eighty papers were retained for the review. We stress that these papers do not represent the entire literature pool in the area. Had we chosen different search terms or used different combinations of terms in our searches, the basis for our review would possibly have been different. The three categories for our analyses were; *socio-demographic factors, immigrant status and ethnicity* and *health and physical characteristics* and we sorted papers by these categories. An additional 61 articles were used to provide background information and possible explanations

There are a number of considerations which impinge on the effect of socio-demographic factors on the prevalence of ASD and/or ID. Firstly, in persons with ASD with ID and those with just ID, the features overlap to some degree. A child with ASD with ID, particularly in the past, may often have been diagnosed with only ID as there have been secular changes in the identification of children with ASD. Secondly, persons who could be diagnosed with ASD without ID are most often able to function independently and may remain undiagnosed in a range of scenarios. Thirdly, the process through which children are assigned a diagnosis of ASD is much more complex than for ID. Whilst elsewhere, the gold standard might be the considered judgement of an expert clinician who had seen many patients with autism [36], in Western Australia, a diagnosis of ASD in a child requires an assessment by a team comprising a paediatrician, psychologist, and speech pathologist. Waiting times for this assessment can be prolonged in the Western Australian public system [37] and even longer in the US. [38] One effect of these considerations is that in some countries a child whose parents are socioeco‐ nomically disadvantaged may oft times be diagnosed with ID when a diagnosis of ASD could

Thirteen articles researched the parental socio-economic status (SES) of children with ASD and/or ID [17, 20, 21, 39-48] and all but two [42, 43] supported a different association of SES with ASD than ID. Children with ASD were more likely to be from higher SES families but

A range of measures of high SES were consistently associated with ASD. In a large telephone interview study in the US, family wealth was used as a measure of SES. [40] The researchers

children with ID were more likely to be from lower SES families.

parents or mothers of children without disability or with a population norm;

least 15 subjects;

child with ASD and/or ID; and

390 Recent Advances in Autism Spectrum Disorders - Volume I

of some of the reported associations.

**3. Socio-demographic factors**

be more appropriate.

**3.1. Socio-economic status**

were assessed as unlikely to lead to bias.

The overall association between high SES and ASD without ID could result from the increased empowerment of parents of high SES to pursue a diagnosis where their children have a milder variant of ASD. [49] In families of low SES, higher functioning children with autistic traits might be informally labelled by family and contemporaries as unusual, difficult or emotionally damaged. In a comparable way, lower functioning children with autistic traits might be formally or informally given a diagnosis of ID. Others have suggested that children of lower SES parents might be more likely to be diagnosed at a later age than those of higher SES and hence not be included in studies of ASD and SES with lower ages of cut-off. [49]

Further evidence of the possible social contributions to the likelihood of an ASD diagnosis was found in a large multi-based national study in the US. [39] Undiagnosed children who met the criteria for ASD had a lower SES than children who had been previously diagnosed. [39] Arealevel SES indicators derived from census data were used in another study where the research‐ ers elucidated that increasing SES and the increasing prevalence of ASD were associated in a dose-response fashion. [39]

King et al. [41] provided evidence that an interaction of social factors was affecting the likelihood of an ASD diagnosis. They examined factors influencing the likelihood of an ASD diagnosis using data on around five million births in Californian cohorts from 1992 to 2000. They found that an interaction between high and low level SES measures influenced the likelihood of an ASD diagnosis. Medi-Cal is a program providing medical assistance to the needy in California and these researchers used family use of Medi-Cal as a binary measure of SES. Property values in the area of a mother's residence were also used as a measure of SES. These researchers reported that children whose families were enrolled with Medi-Cal births and living in wealthier neighbourhoods were two and a half times more likely to receive a diagnosis of ASD than their counterparts living in poorer areas. [41] This could indicate that for parents of limited resources, living in a higher SES neighbourhood had benefits in terms of the likelihood of their child being diagnosed with ASD. Possibly, this results from the parents' increased access to support persons such as paediatricians and child health nurses and to educational programs such as parent classes and interventions for children, compared to that of similar parents in less affluent areas.

In contrast, a Danish study accessing linked population data, used maternal education and parental wealth as a measure of SES and found no association between SES and ASD diagnosis. [42] In neighbouring Sweden, a population-based study published in 2012, used low income, manual occupation and less education as measures of low SES. The researchers concluded that low, not high SES, was a risk factor for ASD. [43] There may be a number of reasons for the differing findings of these studies. The universal health-care and routine screenings offered in Denmark and Sweden may eliminate the ascertainment bias associated with high SES which may exist in other Western countries. [43]

In Australia, a retrospective cohort study, using linked health registries assessed marital status in terms of living with a partner. They reported that at their child's birth, women living with a partner were 35% more likely to have a child with ASD and particularly ASD with ID. [17] On the other hand, a similar Canadian study found that mothers not living with a partner at the time of their child's birth were 19% more likely to have a child with ASD than those mothers

Pre-Existing Differences in Mothers of Children with Autism Spectrum Disorder and/or Intellectual Disability: A Review

http://dx.doi.org/10.5772/54488

393

With ID, women without a partner had increased odds of having a child with ID [17] and particularly mild or moderate ID. [20] Similarly, a cohort study using UK data, concluded that compared to typically developing children, those with early cognitive delay were less likely to have their biological parents living together during the first five years of their lives compared to families with a typically developing child. [55] However, in Finland, the negative association between living with a partner and the odds of ID in the offspring, present in a 1966 birth cohort, was absent in the 1985-6 cohort. [46] The reduction of the association in the second cohort may

have been a reflection of the improved SES of single mothers over the 20 year period.

In most studies, increasing maternal age, sometimes along with increasing paternal age, was associated with ASD. A minority of studies found relationships only with paternal age or found no association with either maternal or paternal age. Contrasting results were reported with ID where teenage mothers were more likely to have children with mild or moderate ID were older mothers and particularly likely to have children with severe or profound ID. Socio-demo‐

All ten studies investigating the association of maternal age with the prevalence of ASD found that advanced maternal age was associated with an increasing prevalence of ASD [17, 29, 47, 56-61] and sometimes ASD without ID. [17, 47] Four of these studies, reported an additional association with paternal age. [17, 56, 58, 61] For instance, a population-based study using data from multiple sites throughout the US, found associations with both maternal and paternal age after adjustment for the other parent's age, birth order and maternal education. [58]

Five of the cited studies specifically reported an association between paternal but not maternal age and ASD in the offspring. [58, 62-65] One of these studies was a small Japanese case-control study of 84 father-child dyads. The researchers reported that advanced parental age was associated with nearly twice the risk of ASD without ID. [65] Another was a population-based Israeli cohort study which used data from a medical registry. [63] The remaining studies used population data from Sweden and another, population data from Denmark. [62, 64] After an adjustment for maternal age, the Swedish researchers identified a linear association of increasing paternal age and the risk of ASD. These researchers commented that if no adjust‐ ment was made for paternal age it would appear as though maternal age, rather than paternal was the risk factor for ASD. They added that paternal age could be a risk factor as generally the male was considered to be the origin of new mutations in the gene pool and their produc‐

who were living with a partner. [54]

graphic and biological explanations are offered.

tion increased with age. [62]

**3.3. Parental age**

By comparison with ASD, low SES was often identified as a risk factor for ID [21, 44-46, 48] and especially mild or moderate ID. [17, 20, 48] One of these studies was a cross-sectional study of over five million children. [48] It concluded that children with mild or moderate ID had an increased risk of exposure to social conditions which were detrimental to their development. [48] Another study examined SES and ID prevalence in the 1966 and 1985-6 Finnish birth cohorts. [46] The researchers concluded that the association of low SES with ID was present in both cohorts. Plausible hypotheses for this persisting association are that there had been no improvements in antenatal and obstetric care in those of lower SES over the twenty years in question or, alternatively, there is a prominent genetic involvement in the aetiology of ID. Another, is that the higher risk of exposure to a developmentally unfavourable environment has persisted over the 20 year interval in the children of mothers of lower SES. [48]

In total, ten studies [20, 21, 23, 41, 45, 47, 50-53] used education alone as a measure of SES. All four of the studies investigating ASD reported positive associations between high maternal education and the risk of ASD in the offspring. Three of these studies were from California and each reported that parents of children with ASD were more educated than the general population. [21, 41, 50] The fourth reported that mothers with more than 16 years education were more than twice as likely to have a child with ASD without ID than mothers of a child with only 12 years education. [47] The relationship was reversed with maternal education and ID where all research ascertained a negative association between high maternal education and the risk of ID in the offspring. For instance, with children with unspecified ID [20, 21, 23, 45, 51, 53] and developmental delay without ASD [24](which may include those with known genetic syndromes), seven studies concluded that their mothers were of a lower educational status. One of these, a population study, established that mothers of children with ID were less likely to have more than 13 years of education. [23]

The association of maternal education with varying levels of ID has been investigated includ‐ ing for severe ID and on the basis that risk factors for Down syndrome differed from those of other forms of ID, children with Down syndrome were excluded. Mothers of children with severe ID were found to be more likely to have a lower educational status than mothers in the general population. [52] Comparable results were found for mothers of children with mild or moderate ID [20, 21] of unknown cause. These mothers had increased odds of a lower educational status than mothers in the general population. One of these studies used Califor‐ nian service agency records and a sample of more than 27 000 mothers of children with mild or moderate ID or severe or profound ID. [21] Less maternal education was also associated with an increased risk of severe or profound ID in the offspring.

#### **3.2. Marital status**

Four papers, describing five studies, examined marital status in relation to the odds of ASD and/or ID. [17, 20, 46, 54] At the time of their child's birth, it is uncertain whether a woman's marital status is associated with her odds of a child with ASD. However, mothers of children with ID were more likely to be without partners.

In Australia, a retrospective cohort study, using linked health registries assessed marital status in terms of living with a partner. They reported that at their child's birth, women living with a partner were 35% more likely to have a child with ASD and particularly ASD with ID. [17] On the other hand, a similar Canadian study found that mothers not living with a partner at the time of their child's birth were 19% more likely to have a child with ASD than those mothers who were living with a partner. [54]

With ID, women without a partner had increased odds of having a child with ID [17] and particularly mild or moderate ID. [20] Similarly, a cohort study using UK data, concluded that compared to typically developing children, those with early cognitive delay were less likely to have their biological parents living together during the first five years of their lives compared to families with a typically developing child. [55] However, in Finland, the negative association between living with a partner and the odds of ID in the offspring, present in a 1966 birth cohort, was absent in the 1985-6 cohort. [46] The reduction of the association in the second cohort may have been a reflection of the improved SES of single mothers over the 20 year period.

#### **3.3. Parental age**

Denmark and Sweden may eliminate the ascertainment bias associated with high SES which

By comparison with ASD, low SES was often identified as a risk factor for ID [21, 44-46, 48] and especially mild or moderate ID. [17, 20, 48] One of these studies was a cross-sectional study of over five million children. [48] It concluded that children with mild or moderate ID had an increased risk of exposure to social conditions which were detrimental to their development. [48] Another study examined SES and ID prevalence in the 1966 and 1985-6 Finnish birth cohorts. [46] The researchers concluded that the association of low SES with ID was present in both cohorts. Plausible hypotheses for this persisting association are that there had been no improvements in antenatal and obstetric care in those of lower SES over the twenty years in question or, alternatively, there is a prominent genetic involvement in the aetiology of ID. Another, is that the higher risk of exposure to a developmentally unfavourable environment

has persisted over the 20 year interval in the children of mothers of lower SES. [48]

In total, ten studies [20, 21, 23, 41, 45, 47, 50-53] used education alone as a measure of SES. All four of the studies investigating ASD reported positive associations between high maternal education and the risk of ASD in the offspring. Three of these studies were from California and each reported that parents of children with ASD were more educated than the general population. [21, 41, 50] The fourth reported that mothers with more than 16 years education were more than twice as likely to have a child with ASD without ID than mothers of a child with only 12 years education. [47] The relationship was reversed with maternal education and ID where all research ascertained a negative association between high maternal education and the risk of ID in the offspring. For instance, with children with unspecified ID [20, 21, 23, 45, 51, 53] and developmental delay without ASD [24](which may include those with known genetic syndromes), seven studies concluded that their mothers were of a lower educational status. One of these, a population study, established that mothers of children with ID were less

The association of maternal education with varying levels of ID has been investigated includ‐ ing for severe ID and on the basis that risk factors for Down syndrome differed from those of other forms of ID, children with Down syndrome were excluded. Mothers of children with severe ID were found to be more likely to have a lower educational status than mothers in the general population. [52] Comparable results were found for mothers of children with mild or moderate ID [20, 21] of unknown cause. These mothers had increased odds of a lower educational status than mothers in the general population. One of these studies used Califor‐ nian service agency records and a sample of more than 27 000 mothers of children with mild or moderate ID or severe or profound ID. [21] Less maternal education was also associated

Four papers, describing five studies, examined marital status in relation to the odds of ASD and/or ID. [17, 20, 46, 54] At the time of their child's birth, it is uncertain whether a woman's marital status is associated with her odds of a child with ASD. However, mothers of children

may exist in other Western countries. [43]

392 Recent Advances in Autism Spectrum Disorders - Volume I

likely to have more than 13 years of education. [23]

with ID were more likely to be without partners.

**3.2. Marital status**

with an increased risk of severe or profound ID in the offspring.

In most studies, increasing maternal age, sometimes along with increasing paternal age, was associated with ASD. A minority of studies found relationships only with paternal age or found no association with either maternal or paternal age. Contrasting results were reported with ID where teenage mothers were more likely to have children with mild or moderate ID were older mothers and particularly likely to have children with severe or profound ID. Socio-demo‐ graphic and biological explanations are offered.

All ten studies investigating the association of maternal age with the prevalence of ASD found that advanced maternal age was associated with an increasing prevalence of ASD [17, 29, 47, 56-61] and sometimes ASD without ID. [17, 47] Four of these studies, reported an additional association with paternal age. [17, 56, 58, 61] For instance, a population-based study using data from multiple sites throughout the US, found associations with both maternal and paternal age after adjustment for the other parent's age, birth order and maternal education. [58]

Five of the cited studies specifically reported an association between paternal but not maternal age and ASD in the offspring. [58, 62-65] One of these studies was a small Japanese case-control study of 84 father-child dyads. The researchers reported that advanced parental age was associated with nearly twice the risk of ASD without ID. [65] Another was a population-based Israeli cohort study which used data from a medical registry. [63] The remaining studies used population data from Sweden and another, population data from Denmark. [62, 64] After an adjustment for maternal age, the Swedish researchers identified a linear association of increasing paternal age and the risk of ASD. These researchers commented that if no adjust‐ ment was made for paternal age it would appear as though maternal age, rather than paternal was the risk factor for ASD. They added that paternal age could be a risk factor as generally the male was considered to be the origin of new mutations in the gene pool and their produc‐ tion increased with age. [62]

By comparison, three studies from Northern Europe, and UK identified that neither of advancing maternal nor advancing paternal age was a risk factor for ASD. [42, 43, 66] One of the studies from Denmark and another from Sweden used linked data from national registries. [42, 43] The third was a much smaller UK study of around 5 000 participants and parents provided data by completing self-reports. As with broader measures of SES, the results from Denmark and Sweden might reflect the model of health service provision in Scandinavia. Moreover, there is evidence that children with ASD are diagnosed later in younger mothers. [67] Thus there may be a bias of ascertainment in some studies where younger children are included. In the UK study, [66] younger mothers may been included more often since they were recruited when pregnant. Further, a diagnosis of ASD was not required for their child but instead, a parent completed the *Social and Communication Disorders Checklist*. In other studies from the US, [41, 47, 58, 59] Canada [29] and Australia, [17, 57] ASD may be underascertained in the children of younger parents, possibly as a result of their lesser confidence to be pro-active in the diagnostic process.

that older nulliparous women with older partners were around three times more likely to have

Pre-Existing Differences in Mothers of Children with Autism Spectrum Disorder and/or Intellectual Disability: A Review

http://dx.doi.org/10.5772/54488

395

However, two studies found other associations between parity and the risk of ASD. The first, a Danish case-control study nested in population data, found no association. [42] The second, a prospective cohort study using linked health data of more than 110 000 mothers in the US, asserted that mothers of parity greater than two were more likely to have a child with ASD than other mothers. [25] Possibly, socio-demographic factors were also operating in this circumstance. In relation to SES and the odds of ASD, it is possible once again that the disparate findings of this same Danish study may have been due to less ascertainment bias which set them apart from other studies in the area. [42] The second study involved nearly 120 000 nurses who were followed via their completion of mailed questionnaires over sixteen years. [25] Hence, all mothers were educated and due to their involvement with nursing, could be expected, on average, to have more knowledge of ASD than other mothers. Further, parity was assessed as a binary variable with the two values of greater than two and less than or equal to two. Commonly, other studies have defined parity as either a continuous variable or one with more than two possible values and this difference might account for variations in study

Mothers of higher parity had increased odds of having a child with mild or moderate ID. [17, 20, 21] One of the research groups concluded that fourth or subsequent children had an increased risk of mild-moderate ID. [20] A Finnish study of two birth cohorts, twenty years apart, found that high parity persisted as a risk factor for ID over time. [46] A large cohort study compared the parity of the mothers of Californian children with ID to the parity of mothers of typically developing children born between 1987 and 1994. [21] These researchers reported that mothers of parity of three or more were 30-50% more likely to have a child with mild or moderate ID or unspecified ID. [21] Both this study and another Californian study reported that mothers of children with severe or profound ID had an elevated but not significantly increased parity compared to mothers of typically developing children. [21, 52]

Socio-demographic factors often operate quite differently for ASD and ID. For example, high parental SES was positively associated with the risk of ASD and negatively associated with the risk of ID in the offspring. Marital status, as defined by living with a partner, has different associations. At the time of their child's birth, there was no consistent association of marital status with mothers of a child with ASD compared to the mothers of typically developing children. On the other hand, mothers of a child with ID were less likely to be living with a partner than mothers of typically developing children. Parity appeared to have reverse associations for ASD and ID. Compared to mothers of typically developing children, mothers of low parity were more likely to have a child with ASD and mothers of high parity were more likely to have a child with ID. Similar patterns exist for maternal age. Mothers of an advanced age were more likely to have a child with ASD than mothers of typically developing children. In contrast, mothers of a younger age were more likely to have a child with ID than mothers

a child with ASD. [58]

findings.

**3.5. Summary**

Maternal age had a dual association with ID of unknown cause. Firstly, teenage mothers were more likely to have children with mild or moderate ID. [17, 20, 21] Secondly, older women were more likely to have children with severe or profound ID. [21, 68] The results of a Finnish cohort study which investigated ID of both known and unknown cause [46] was discounted because of the inclusion of ID of known cause. With Down syndrome, the most common cause of ID, it is known that the risk increases very abruptly with advancing maternal age. [69] This might explain the researchers' finding of an association between increased maternal age and ID in the offspring seen in the 1966 birth cohort. [46] The finding that the association no longer existed in 1985-6 cohort may have been because of the introduction or increased uptake of prenatal screening for Down syndrome.

The association of parental and particularly maternal age with ASD and/or ID suggests that both social and biological forces are operating. Younger parents may find a diagnosis of ASD more difficult to obtain for their children because of inexperience and navigational require‐ ments of local systems. Thus, some of the ID diagnoses of their children may be undiagnosed cases of ASD. Further, the excess of older mothers of children with ASD and to a lesser extent ID may result from increased de novo mutations in older women and their partners [70] or the increase of epigenetic mechanisms which are associated with ageing. [71]

#### **3.4. Parity**

Parity describes the number of live-born children and stillbirths at more than 20 weeks gestation of a woman. [72] Two strong relationships of low parity with ASD and high parity with ID have been demonstrated in the majority of studies.

In women of lower parity, the risks of ASD, [29, 41, 73] ASD with ID [17] and ASD without ID [17, 74] were found to be increased in a number of studies. One of these was a Canadian cohort study using linked data-bases and with nearly 1 000 case mothers. [29] The authors identified that nulliparous women (that is women having their first child) were at the greatest risk of having a child with ASD. Moreover, a national, population-based study in the US reported that older nulliparous women with older partners were around three times more likely to have a child with ASD. [58]

However, two studies found other associations between parity and the risk of ASD. The first, a Danish case-control study nested in population data, found no association. [42] The second, a prospective cohort study using linked health data of more than 110 000 mothers in the US, asserted that mothers of parity greater than two were more likely to have a child with ASD than other mothers. [25] Possibly, socio-demographic factors were also operating in this circumstance. In relation to SES and the odds of ASD, it is possible once again that the disparate findings of this same Danish study may have been due to less ascertainment bias which set them apart from other studies in the area. [42] The second study involved nearly 120 000 nurses who were followed via their completion of mailed questionnaires over sixteen years. [25] Hence, all mothers were educated and due to their involvement with nursing, could be expected, on average, to have more knowledge of ASD than other mothers. Further, parity was assessed as a binary variable with the two values of greater than two and less than or equal to two. Commonly, other studies have defined parity as either a continuous variable or one with more than two possible values and this difference might account for variations in study findings.

Mothers of higher parity had increased odds of having a child with mild or moderate ID. [17, 20, 21] One of the research groups concluded that fourth or subsequent children had an increased risk of mild-moderate ID. [20] A Finnish study of two birth cohorts, twenty years apart, found that high parity persisted as a risk factor for ID over time. [46] A large cohort study compared the parity of the mothers of Californian children with ID to the parity of mothers of typically developing children born between 1987 and 1994. [21] These researchers reported that mothers of parity of three or more were 30-50% more likely to have a child with mild or moderate ID or unspecified ID. [21] Both this study and another Californian study reported that mothers of children with severe or profound ID had an elevated but not significantly increased parity compared to mothers of typically developing children. [21, 52]

#### **3.5. Summary**

By comparison, three studies from Northern Europe, and UK identified that neither of advancing maternal nor advancing paternal age was a risk factor for ASD. [42, 43, 66] One of the studies from Denmark and another from Sweden used linked data from national registries. [42, 43] The third was a much smaller UK study of around 5 000 participants and parents provided data by completing self-reports. As with broader measures of SES, the results from Denmark and Sweden might reflect the model of health service provision in Scandinavia. Moreover, there is evidence that children with ASD are diagnosed later in younger mothers. [67] Thus there may be a bias of ascertainment in some studies where younger children are included. In the UK study, [66] younger mothers may been included more often since they were recruited when pregnant. Further, a diagnosis of ASD was not required for their child but instead, a parent completed the *Social and Communication Disorders Checklist*. In other studies from the US, [41, 47, 58, 59] Canada [29] and Australia, [17, 57] ASD may be underascertained in the children of younger parents, possibly as a result of their lesser confidence

Maternal age had a dual association with ID of unknown cause. Firstly, teenage mothers were more likely to have children with mild or moderate ID. [17, 20, 21] Secondly, older women were more likely to have children with severe or profound ID. [21, 68] The results of a Finnish cohort study which investigated ID of both known and unknown cause [46] was discounted because of the inclusion of ID of known cause. With Down syndrome, the most common cause of ID, it is known that the risk increases very abruptly with advancing maternal age. [69] This might explain the researchers' finding of an association between increased maternal age and ID in the offspring seen in the 1966 birth cohort. [46] The finding that the association no longer existed in 1985-6 cohort may have been because of the introduction or increased uptake of

The association of parental and particularly maternal age with ASD and/or ID suggests that both social and biological forces are operating. Younger parents may find a diagnosis of ASD more difficult to obtain for their children because of inexperience and navigational require‐ ments of local systems. Thus, some of the ID diagnoses of their children may be undiagnosed cases of ASD. Further, the excess of older mothers of children with ASD and to a lesser extent ID may result from increased de novo mutations in older women and their partners [70] or the

Parity describes the number of live-born children and stillbirths at more than 20 weeks gestation of a woman. [72] Two strong relationships of low parity with ASD and high parity

In women of lower parity, the risks of ASD, [29, 41, 73] ASD with ID [17] and ASD without ID [17, 74] were found to be increased in a number of studies. One of these was a Canadian cohort study using linked data-bases and with nearly 1 000 case mothers. [29] The authors identified that nulliparous women (that is women having their first child) were at the greatest risk of having a child with ASD. Moreover, a national, population-based study in the US reported

increase of epigenetic mechanisms which are associated with ageing. [71]

with ID have been demonstrated in the majority of studies.

to be pro-active in the diagnostic process.

394 Recent Advances in Autism Spectrum Disorders - Volume I

prenatal screening for Down syndrome.

**3.4. Parity**

Socio-demographic factors often operate quite differently for ASD and ID. For example, high parental SES was positively associated with the risk of ASD and negatively associated with the risk of ID in the offspring. Marital status, as defined by living with a partner, has different associations. At the time of their child's birth, there was no consistent association of marital status with mothers of a child with ASD compared to the mothers of typically developing children. On the other hand, mothers of a child with ID were less likely to be living with a partner than mothers of typically developing children. Parity appeared to have reverse associations for ASD and ID. Compared to mothers of typically developing children, mothers of low parity were more likely to have a child with ASD and mothers of high parity were more likely to have a child with ID. Similar patterns exist for maternal age. Mothers of an advanced age were more likely to have a child with ASD than mothers of typically developing children. In contrast, mothers of a younger age were more likely to have a child with ID than mothers of typically developing children. However, an additional association exists with older mothers being also more likely to have a child with severe ID.

ASD than other immigrant mothers. [57] A Western Australian study, using linked population data, also found that immigrant mothers from South-East or North-East Asia were at increased risk of having a child with ASD with ID. [17] A similar situation was described in Sweden where immigrant mothers from East Asia were more than three times as likely to have a child

Pre-Existing Differences in Mothers of Children with Autism Spectrum Disorder and/or Intellectual Disability: A Review

http://dx.doi.org/10.5772/54488

397

Black immigrant mothers and immigrant mothers from developing countries were also found to be more likely to have a child with ASD compared to other immigrant mothers. One study from the UK [75] and another from Sweden [76] reported that black immigrant mothers [75] and immigrant mothers from sub-Saharan Africa [76] were much more likely to have a child with ASD compared to non-immigrant mothers. Further, a small Swedish case-control study compared the prevalence of autistic disorder and pervasive development disorder not otherwise specified (PDDNOS) in black African children with at least one parent born in Somali to the prevalence in children without a Somali background. [77] The researchers reported that these 17 black mothers were from three to four times more likely to have a child with ASD

There is evidence that 'the intensity of the mother's skin colour' is related to her risk of having a child with ASD. A Swedish study compared the risk of ASD in the children of immigrants from each of North, East and other parts of Africa. [22] The mothers from North Africa were predominantly Moroccan and hence were probably fairer than the other two groups of mothers. For example, the East African group was predominantly from Somalia and Ethiopia while the ethnicity of the group from other parts of Africa was not described. The risk of ASD in the North African group was elevated (1.5) but not significantly higher than that of nonimmigrant parents. On the other hand, the risk in the East African mothers and mothers from

Immigrant mothers from distant countries and those who emigrated during pregnancy were more likely to have a child with ASD than other immigrant mothers. For instance, researchers from the UK and Denmark found that immigrant mothers born outside of Europe were more likely to have a child with ASD. [62, 75] Similarly, a Swedish study found that immigrant mothers who were not from either of the US or Europe were nearly three times as likely to have a child with ASD compared to mothers from Nordic countries. [30] Another Swedish study ascertained that immigrant mothers who emigrated during pregnancy were even more

There is evidence that immigrant mothers are at different risks of ASD without ID and ASD with ID. Two Swedish studies found that immigrant mothers, excepting those from neigh‐ bouring Northern Europe, were less likely to have a child with ASD without ID [22] and Asperger syndrome [76] compared to non-immigrant mothers. One of these studies, along with an Australian study, reported that immigrant mothers were more likely to have a child with ASD with ID. [17, 22] In addition, the Swedish study found that the African immigrant mothers were more likely to have a child with ASD with ID compared to non-immigrant mothers. [22] Similar results were found in a small Swedish case-control study, where all

likely to have a child with ASD than mothers who emigrated at other times. [22]

seventeen of the Somali children with autism presented with ASD with ID. [77]

compared to the mothers without a Somali background. [77]

other parts of Africa of having a child with ASD was 1.9 and 3.5. [22]

with ASD. [76]

An under-ascertainment of ASD due to social factors and, to a lesser extent an over-ascertain‐ ment of ID could be contributing to the socioeconomic effects seen with ASD and ID. For instance, in terms of the severity of ASD, researchers in California, with birth cohorts from 1992 to 2000, divided the children with ASD into two groups of equal size where the less severe group comprised children in the top 50% of cases according to level of functioning and the most severe group was the lower 50%. [41] They found that the children from the less severe group were more often found in neighbourhoods which housed wealthier and more educated individuals. Conversely, the same researchers reported that where low SES was measured by a Medi-Cal payment for the birth, the ratio of more severe to less severe cases was always greater than one. The researchers' interpretation was that the most difficult to diagnose cases of ASD, that is the less severely affected, were under-ascertained in lower SES populations. [41]

The association of high SES with ASD also might be compounded by some of the characteristics known to be related to mothers of children with ASD. Older women with the support of a partner and with fewer children would seem more likely to achieve a more complex diagnosis requiring more assessments for their child than younger single mothers. Socio-demographic associations with ASD in most Western countries do not appear to operate as strongly and might even be absent in some Northern European countries. This might be due to a different social welfare structure in this region and specifically related to the universal screening for developmental disability. In addition to these and other social factors which could bias ascertainment, biological factors may be operating with older parents.
