National screening for Egyptian children aged 1 year up to 12 years at high risk of Autism and its determinants: a step for determining what ASD surveillance needs

According tothe Diagnostic and Statistical Manual of Mental Disorders (5th edition) (DSM-5) criteria, autism spectrum disorder (ASD) is defined as a group of neurodevelopmental disorders characterized by early-onset and persistent impaired social interaction, communication, and restricted/repetitive behavior and interest [22, 43, 96]. Having a child with ASD is accompanied by a great burden of care on family members; physically, emotionally, and financially [7, 57].

The Global Burden of Disease Study estimated ASD prevalence and years living with disability among the common six prevalent developmental disabilities in children under the age of five years [73]. The prevalence of ASD has increased considerably in the last two decades, reaching around 1–2% of the general population due to greater recognition and knowledge of the disease [12].

Around 52 million children with ASD were recorded worldwide in 2010, with a prevalence of 7.6 per 1000 [88]. In 2018, the Centers for Disease Control and Prevention (CDC) reported that nearly 1 in 59 children had ASD. In 2020, that number increased to 1 in 44 children [19]. The most recent data from Europe and the United States show a significant increase in the measured prevalence of autism over the last two decades from 0.48 to 3.13 percent [66]. However, research on ASD prevalence in Arab countries was limited to high-income countries; Systematic meta-analysis in these countries revealed wide variations, Oman, the United Arab of Emirates, Saudi Arabia, Bahrain, Kuwait, and Qatar [1, 8]. Meanwhile, there is a shortage of data concerning ASD prevalence in developing Arab countries. Generally speaking, the great variation in the prevalence of ASD estimates that was found in Arab countries was due to the variation in the study methods, age of participants, case ascertainment approaches, and diagnostic tools used as there is no standardized method for screening of ASD until now which is considered a challenging situation. Due to the same reasons, studies in Egypt estimated great variability of ASD prevalence ranging from 5.4/1000 [102] to 33.6% [85]. At the same time, studies in Egypt were limited to, confined areas, facility-based, and with a limited sample size [41, 102].

Till now, the causes and risk factors of ASD are not well known with many operational risk factors; prenatal, perinatal, and environmental [87, 97]. Even though, there is growing evidence that ASD has become a public health problem in many countries. Such findings indicated that autism has to be perceived as an important problem that deserves to be in the light spot for policymakers to allow delivery of targeted services, especially in developing countries where access to therapeutic effective services is either limited or not existed. Early detection of ASD is crucial for prompt and early intervention to achieve significant improvement in behavioral performance and the intellectual abilities of children with ASD.

This study aimed at providing a national screening estimate of children who are at high risk of ASD among Egyptian children aged 1 to 12 years with integration of their developmental level based on reliable screening tools used during two phases. The study identified the possible determinants of the at high risk of ASD that are related to sociodemographic, epidemiological, maternal, and child health status.

Table 1 shows the socio-demographic characteristics and perinatal problems of the study population. The parents or the caregivers of 41,640 children in the age range 1–12 years were participated in the study. Children were slightly insignificantly higher in rural than urban communities. They were equally distributed among social classes. They were distributed among the randomly selected governorates according to their population size. The surveyed boys were insignificantly slightly higher than girls (51.5% versus 48.5% respectively). Half of the children were in the age group 6 to 12 years (48.8%). The least percentage of children were in the age group 1- < 3 years (20.1%). 95.8% of children above 6 years went to school. The mean age at which mothers gave birth was 26.06 ± 6.13 years and their current mean age was 32.04 ± 6.81 years. Regarding education levels, almost half of mothers and fathers completed their higher school degrees (44.7% and 44.2% respectively). Most of the mothers were unemployed (84.8%). Houses without mothers were 0.7% versus 4.6% without fathers. Among the perinatal conditions, the presence of a history of neonatal jaundice after birth was the most prevalent (26.5%) followed by a history of difficult labor (14.6%). 0.9% of houses had mothers who were physically or mentally disabled and 4.3% had twins. History of LBW was reported among 4.4% versus 1% with a history of premature delivery.

Age and sex-specific estimates of ASD appeared in Figs. 1 and 2 respectively. Group-specific prevalence was highest for the age group 5- < 6 years (6.4%, 95% CI: 5.7%–7.2%) followed by age group 6- < 7 (5.7%, 95% CI: 5.0%–6.5%) as shown in Fig. 1.

The prevalence of ASD among boys was almost double that of the girls for these two specific age groups (COR = 1.70, 95% CI: 1.32%–2.19% & COR = 2.06, 95% CI: 1.54%–4.88% respectively) as shown in Fig. 2.

Summing the prevalence of the at high risk of ASD among children aged 1–12 years was 3.3% (95% CI: 3.1%–3.5%). The prevalence of the at high risk of ASD among boys aged 1–12 years was 4.1% (95% CI: 3.4%–4.8%) versus 2.4% (95% CI: 1.7%–3.3%) among girls in the same age group. Having a child at high risk of ASD was likely to be found among boys one and three-quarters times that of girls in the age range 1–12 years (COR 1.72, 95% CI: 1.53%–1.92%).

Table 2 shows the odds of having features of autism among surveyed children according to the socio-demographic and child characteristics. Having at high risk of ASD was higher among children belonging to the middle class than belonging to either the high and low classes (COR 1.53, 95% CI: 1.34–1.73 & COR 1.72, 95% CI: 1.51–1.97 respectively). Living in urban communities was associated with significantly increased odds of at high risk for ASD than living in rural (COR = 1.71, 95% CI: 1.53–1.91). Living in Lower Egypt or in Frontiers significantly decreased the odds of at high risk of ASD than living in cities by more than 60% (COR = 0.39, CI: 0.34–0.44 & COR = 0.36, CI: 0.29–0.45 respectively).

Children aged 3 – < 6 years were the most likely to be at high risk of ASD with the highest prevalence (4.4%) with significantly twice and three quarters more odds than those of the age group 1- < 3 years (COR = 2.7, CI: 2.29–3.34) and one and a half more odds than children aged 6–12 years (COR = 1.46, CI: 1.21–1.52).

Non-school attendance was significantly five times more among children with at high risk of ASD (COR = 5.2, 95% CI: 4.43–6.17) than non-autistic children.

Concerning the odds of having at high risk of ASD among surveyed children according to the maternal and paternal characteristics (Table 3), the age of mothers who gave birth above 35 years was associated with significantly increased odds of ASD than mothers giving birth below 18 years by one and a half-time (COR = 1.57, 95% CI: 1.13–2.18).

Children with mothers or fathers who had higher education were less likely to have at high risk of ASD, especially those who had a college or greater education level with a fewer odds of a range from 25 to 39% than lower grades of education. The influence of higher paternal education was higher than that of maternal education. Mothers’ work or current age did not show any association with having at high risk of ASD.

Whereas living without mothers was associated with two times increased odds for at high risk of ASD (COR = 2.04, 95% CI: 1.29–3.23) than living with mothers. Living without fathers carries significantly one and a half odds for at high risk of ASD than the presence of fathers (COR = 1.61, 95% CI: 1.30–1.98).

The odds of having at high risk of ASD among surveyed children according to all the studied medical perinatal history problems were shown in Table 4. All the studied medical histories for both mothers and children were risk factors for at high risk of ASD. The odds for at high risk of ASD are largely affected by the type of medical perinatal history.

Mothers with apparent physical or mental disabilities carried the highest odds for the at high risk of ASD among the other maternal problems (COR = 7.77, 95% CI: 5.9–10.1). Whereas, children with a history of cyanosis after birth (COR = 6.23, 95% CI: 4.94–7.85) or convulsions (COR = 7.8, 95% CI: 6.40–9.51) or had a history of meningitis after birth (COR = 4.50, 95% CI: 3.32–6.12) were the most of postnatal child problems for carrying the highest odds for the at high risk of ASD than being a healthy child.

The odds of presenting features of Autism among surveyed children according to sex were shown in Table 5. Three presenting features of Autism were significantly observed to be higher among boys with at high risk of ASD than among girls. Behavioral/Emotional expression problems were significantly two times higher among boys with at high risk of ASD than among girls with at high risk of ASD (COR = 1.96, CI: 1.32–2.91). Whereas, Reduced peer interaction and insistence of children to play with the same thing the same way were significantly more than one and a half times higher among boys with at high risk of ASD than among girls with at high risk of ASD (COR = 1.64, CI: 1.22–2.21 & COR = 1.61, CI: 1.17- 2.19). The odds of having poor to limited comprehension/Limited language/nonverbal, speech difficulties, learning, and academic performance difficulties were insignificantly higher among boys than girls.

Twenty four significant variables associated with children who screened + ve for at high risk of ASD versus those who screened negative of the healthy population were entered into a multivariate logistic regression using the enter selection procedure (Table 6). The odds of being at high risk of autism increased significantly with being males by almost one and three-quarters times that being females (AOR = 1.70, 95%, CI:1.51–1.92).

The neonatal problems related predictors for at high risk of ASD in order were: Children with a history of any convulsions increased the odds of autism by almost four times compared to those not having convulsions (AOR = 3.67; 95% CI: 2.8–4.8), children with a history of cyanosis after birth (AOR = 1.87; 95% CI: 1.35–2.59), history of LBW babies (AOR = 1.53; 95% CI: 1.23–1.89). Children kept in an incubator for more than two days (AOR = 1.47; 95% CI: 1.22–1.76) were associated with increased odds of being at high risk of autism by nearly one and a half times compared to healthy and full-term born children.

Maternal predictors were having either a history of health problems during pregnancy or a history of difficult labor. They increased the odds to be at high risk of autism by one and half times compared to healthy mothers (AOR = 1.74; 95%, CI: 1.45–2.10 & AOR = 1.74; 95%, CI:1.45–2.10). Maternal factors are not age-specific predictors.

The protective factors that decrease the odds to be at high risk of autism included: being resident in a rural locality with a decrease in the odds to be at high risk of autism by 15% compared to being resident in urban communities (AOR = 0.84, 95%, CI: 0.73 – 0.98), maternal and paternal education with university or above degree decrease the odds to be at high risk of autism by approximately 40% compared of being with less degree of education (AOR = 0.66, 95%, CI: 0.52 – 0.85 & AOR = 0.64, 95%, CI: 0.51– 0.82 respectively). Being resident in frontiers or in Lower Egypt decreases the odds of being at high risk of autism by almost 50% compared to being resident in cities (AOR = 1.5, 95%, CI: 1.09 – 2.06).

ASD is a complex developmental disorder that causes a burden to affected children and their families. Early detection and intervention lead to a better prognosis with decreased psychological and economic burdens on the families and society. According to Leo Kanner, the first prevalence estimate of ASD was reported in 1966 to be 4.5 per 10,000 people among children aged 8 to 10 in the United States [51]. Meanwhile, there is a consensus perception about the increase in ASD prevalence worldwide over the last decade. This perception may be attributed to the increased public awareness and better diagnostic services [4], earlier autism diagnosis, and/ or the discovery of therapeutic targets [88]. Accordingly, the current study focused on screening to detect the at high risk of autism with their risk factors. This step is important at the primary health care level to detect suspicious cases for early management. The current study was both a community and facility-based screening study for detecting children at high risk for ASD and determining the risk factors among 41,640 children aged 1–12 years.

There is a growing increase in ASD prevalence worldwide. However, the prevalence is variable in different countries; in the USA, 1 in 59 children had ASD [12] and in the UK, the prevalence of ASD was 1.1% [102]. Few researchers have looked at the national prevalence of ASD among Arab countries as estimated by a recent systematic scoping review [2] with the reported lesser prevalence of ASD than that reported among the developed countries. In Arab countries, it is reported that there is a wide range in prevalence of ASD from 0.014 to 4.7% [2]. The estimate was 2.9 per 1000 in the United Arab Emirates [102], 1.4 per 10,000 in Oman, 4.3 per 10,000 in Bahrain, and 1/167 in Saudi Arabia [5, 81]. The estimated variations in ASD prevalence between countries were mainly attributed to the methodological tools used for the autism diagnosis. It is well known that the confirmation of ASD diagnosis is usually done by specialized clinicians using both DSM5 (Diagnostic and Statistical Manual of Mental Disorders (DSM-5), [23] and CARS [84]. Accordingly, the terminology used for defining autism prevalence should be very precise according to the used tool for diagnosis. Another contributing factor to the detected variation was the targeted age group differences between studies as well as the level at which the study was conducted; whether at the community or facility level.

Few studies were done in Egypt to detect either the prevalence of the at high risk of Autism or ASD confirmed cases; mostly clinic-based, typically with small sample sizes, and were not representative of the whole country. Although ASD is frequently underdiagnosed or misdiagnosed in Egypt [2], existing studies were suggestive that ASD is a prevalent disorder in Egypt. The majority of prevalence studies in Egypt included samples from specialized units or hospitals and were related to the confined age group as a target group. The main objective of the current study was reporting the real situation of children at high risk of ASD as a national screening estimate through two screening phases. The first phase was a community-based one through using Vineland Adaptive Behavior Scales to detect children suspicious for autism. The second screening phase was a facility-based for reporting the prevalence of the high risk of ASD among the referred suspicious children aged 1–12 years using the M-CHAT for children aged 1–3 years, GARS-2 for children 3–12 years. In addition, Denver II Developmental screening test for children up to 6 years was used to confirm developmental delay detected in phase I screening and documenting possible developmental regression that is affected because of autism. The current study estimated the prevalence of the at high risk for autism to be 3.3% (95% CI: 3.1%–3.5%). This estimate was in accordance with two facility-based Egyptian studies. The first study investigated 500 children aged from 3 to 12 years in one Egyptian governorate according to GARS-2 and showed the prevalence of ASD to be 3.4% [11]. The other Egyptian screening study was done in kindergartens on 3722 preschool children and estimated that 2.8% of children were at high risk for ASD according to Modified Check List for Toddlers/Revised (M-CHAT-R) [102].

The current study reported that the highest prevalence was for the age group 5- < 6 years (6.4%, 95% CI: 5.7%–7.2%) followed by age group 6- < 7 years (5.7%, 95% CI: 5.0%–6.5%). The study revealed also that children aged 3 – < 6 years were the most likely to be diagnosed with ASD (4.4%). This was in accordance with another study which found that the most common age of presentation was between 2–5 years of age [103]. This finding supports the crucial recommendation for the early detection of ASD in preschool children [49, 54]. Unfortunately, the majority of children with ASD do not obtain a diagnosis until they reach school age. The fact that many physicians- especially in rural communities- have insufficient awareness of the diagnosis of ASD resulted in a delay in referral of the at-risk children to ASD specialists, another reason was the lack of awareness of mothers about the developmental milestones until their children reach school age [86]. This values the importance of screening at the community level to provide a real situation for ASD problems.

The present study's finding of having that ASD among boys was one and three-quarters times more than that of girls in the age range 1–12 years (OR 1.72, 95% CI: 1.53%–1.92%) was in parallel with many studies that revealed a higher frequency of ASD among males than females with a ratio of 2:1 [3], 3:1 [11, 51], or 4:1 [50, 67]. Results of different studies have been contradictory. When screening the entire population using gold standard assessments, current estimates suggest around 3 males receive an autism diagnosis for every female. However, in clinical samples who have already received an autism diagnosis, that ratio is higher at over four males to each female [50]. In individuals with intellectual disability, the ratio is closer to 2:1 [101].

One of the first signs of ASD is delayed speech. However, the presence of speech before the age of five is the most powerful predictor of better improvement of ASD symptoms [69]. The odds of having speech difficulties, learning, and academic performance difficulties were insignificantly higher among boys than girls in the current study (COR = 1.37, CI: 0.31–1.71). Our results reported three presenting features of autism to be significantly higher among the at-risk boys than among the at-risk girls for ASD. Behavioral and emotional expression problems were significantly associated with nearly two times higher odds of ASD risk among boys than among girls (COR = 1.96, CI: 1.32–2.91). Whereas, reduced peer interaction and insistence of children to play with the same thing the same way carried significantly more than one and a half times higher odds among boys than among girls with at risk of ASD (COR = 1.64, CI: 1.22–2.21 & COR = 1.61, CI: 1.17- 2.19). Our study comes in agreement with Yousef and his colleagues [102] who reported behavioral disturbance, significant developmental and speech delay among ASD patients, also with another Egyptian study that was done by El-Baz and his colleagues who proved that ASD is associated with developmental delay, speech defect, psychological disorders, and intellectual and learning disabilities [28, 40, 45, 102]. Alshaban and his coworkers [8] reported that 75.1% of children with ASD had language delays for words, 91.4% for phrase speaking, and 19.4% had developmental regression. They also reported that 19.4% of children had persistent deficiencies in expressive language and 14.0% in peer interactions.

There have been not enough evaluations on sex differences by ASD end phenotypes to date. However, some researchers reported that males with ASD are more likely than females to suffer from repetitive behaviors and have limited interests [17, 39, 78]. The odds of having poor to limited comprehension or limited language or nonverbal difficulties were insignificantly higher among boys than girls in the present study. This finding was also similar to other researchers who found no significant differences between males and females with ASD in communication, comprehension difficulties, or language skills [71, 89, 91].

ASD risk factors have yet to be properly defined [102]. The current study identified the determinants of children who were at high risk for ASD that are related to socio-demographic, epidemiological, maternal, and child health status. Cultural aspects of ASD remain understudied. Saudi Arabia and Lebanon provided the majority of the studies and declared no association between ASD and socioeconomic status [25, 32]. Investigating the association between parental socioeconomic status and risk of ASD is important to be clarified as it could provide insight into the etiology of ASDs, which is still largely not understood [75]. Some investigators have found a strong link between ASD and urbanity (i.e., a higher risk of autism in urban than rural districts) [48, 77, 98]. Our results revealed the same and reported that living in Lower Egypt or in Frontiers significantly decreased the odds at risk of ASD than living in urban cities by more than 60% (OR = 0.39, CI: 0.34–0.44 & OR = 0.36, CI: 0.29–0.45 respectively). Moreover, the at-risk of ASD was significantly higher among the middle class than in both the high and low classes (OR 1.53, 95% CI: 1.34–1.73 & OR 1.72, 95% CI: 1.51–1.97 respectively). Many studies implicated the role of the socioeconomic status of parents in the early diagnosis of ASD as a low socioeconomic class may delay taking medical advice and this could lead to the child not being probably diagnosed [53, 54]. However, in the current study the association between the middle class and the risk of ASD could be attributed to the lack of functional role of the parents especially mothers. The majority of mothers belonging to the middle class are working for a long time, seldom communicating with their children (play, talk or teach things) being overwhelmed by the home duties, besides their limited awareness about children's developmental milestones. Most of children belonging to the middle class have been constrained for screens for prolonged periods of time. A recent study showed that longer durations of screen time among 1-year-old boys was significantly associated with ASD at 3 years old [47]. On the other hand, women belonging to the low social class leave their children to play outside the home with their peers or mates which had a positive impact on the improvement of their cognitive and social skills. Whereas, children belonging to the high class often regularly attend a preschool education program that enhances children's communication skills.

Systematic review studies revealed that advanced paternal and maternal age has been associated with the risk for autism [52, 55, 74].

When looking at the odds of having ASD, mothers who were over 35 years old at the time of birth carried one-and-a-half- higher odds of risk for ASD compared to mothers who were under 18 years old (OR = 1.57, 95 percent CI: 1.13–2.18) in the current study.

The present study revealed that children with mothers or fathers who had higher education were less likely to be associated with a risk of ASD, especially those who had a college or greater education level with a fewer odds of a ranging from 25 to 39% than lower grades of education indicating that education serves as a protective factor. This could be supported by many Egyptian studies which concluded that awareness of the mothers regarding developmental milestones [27, 64, 79] and their children feeding practices [61] increased with the attainment of maternal education.

Unfortunately, this finding was in contrast with other Egyptian studies for which the majority of ASD children's moms were university graduates [29, 102], A Saudi Arabian study discovered also a high educational level of ASD children's moms [9]. Such variability in the mentioned studies could be attributed to the selection criteria with the exclusion of illiterate parents and those of a low education level from participating due to their negligence in filling the checklist questionnaire. These studies declared their selection as one of their studies’ limitations.

Until now no obvious causes for ASD could be fully understood, no single factor is responsible for the development of ASD but some multifactorial determinants seem to operate in the development of this disorder. On looking up predictors of the at high risk for ASD, prenatal and postnatal insults showed a significant impact. Children with a history of convulsions were associated with increased odds of being at high risk for autism by almost four times compared to those not having convulsions (AOR = 3.67; 95% CI: 2.8–4.8). It has been estimated that between 15 to 47% of people with ASD have epilepsy in early childhood [46].

Meanwhile, children with a history of cyanosis after birth or with a history of LBW or those who were kept in an incubator for more than two days or with a history of LBW babies were associated with increased odds to be at high risk for ASD by nearly one and a half times compared to healthy and full-term born children. The history of maternal health problems during pregnancy or difficult labor was associated with increased odds to be at high risk of ASD by one and half times compared to healthy mothers (AOR = 1.74; 95%, CI:1.45–2.10 & AOR = 1.74; 95%, CI:1.45–2.10). This finding supports other studies that affirmed the significant implication of perinatal insults for the development of ASD [11, 14]. Our study supports the recommendation that policymakers have to be concerned with the provision of services mainly during the perinatal period. In Egypt, community and health facilities based interventions had a great impact on reducing maternal and neonatal adverse outcomes [65, 58].

Our study has some important strengths. To the best of our knowledge, this study is considered to be the first study to investigate the prevalence and predictors for children who are at high risk of ASD at a national level with a very large sample size of Egyptian children aged 1–12 years and as a representative to Egypt. Accordingly, we could make it easy to generalize our findings. The use of two phases for screening using reliable screening tools is a strength. Our study also was the first that determined sex-specific estimates and predictors of ASD.

The lack of studying the environmental variability in the demographic status of the participating children is a weakness of this study. Environmental factors such as industrially emitted heavy metals, chemicals, bacterial, and viral infections may all play a role in ASD. The documented impact of environmental education and awareness-raising at the community level in Egypt in having a profound effect on knowledge attainment was proved [62, 63]. Out of limitation was studying the influence of childhood malnutrition which may be risk factors for ASD, especially the impact of school feeding programs on improving children’s growth and development was documented also in Egypt [60, 80]. The last-mentioned two factors are limited to be investigated during screening studies because of the long time required for filling the related questionnaires.

In this paper, we focused on the whole community to determine the prevalence of children who are at high risk of ASD among 41,640 children aged 1 to 12 years in Egypt through community-based national screening. This study was conducted among 8 Governorates representing all geographic regions of Egypt. The survey highlighted the most important contribution to the need for community-driven data to detect the risk and the protective determinants of ASD in general and among sex which when neglected can lead to a severe degree of ASD that could impact not only the life of the children but the whole family. Most of the detected determinants were preventable factors indicating that the educational and rehabilitation fields are far beyond the medical ones. Thus identifying the prevalence and the risk factors for the high risk of ASD is a step toward implementing surveillance for the at-risk of autism which is drastically needed for creating awareness among health care workers and caregivers. Neglecting to study the extent of the problem of the spread of autism makes more burdens on the families and the governments and helps in increasing the rate of disabilities.

The study findings necessitate prompt actions of local health and educational systems to be appropriately equipped to support affected children and their families optimally. Local health units should be provided with multidisciplinary teams including developmental and behavioral pediatrician, psychiatrist, psychologist, speech therapist, physiotherapist and social worker for early detection, proper diagnosis and management of these children. In addition, efforts must be taken to reduce the uncertainty around the proper estimates of this lifelong disorder. These efforts may include; screening of all children for ASD at ages 9, 18 and 24 months, along with regular developmental surveillance during vaccination visits, screening of preschool and school age children for ASD on admission to kindergarten and primary schools and consistent courses for primary physicians, school teachers, and kindergarten supervisors about early warning signs of ASD.