TABLE 2.
Outcomes of included studies.
| Study | Analysis type, Outcome measures | Covariates | Summary of Findings | Results (unadjusted OR or equivalent) | Results (adjusted OR or equivalent) | Quality rating* |
| Sociodemographic | ||||||
| Abdullahi et al. (29) Western Australia, Australia | Logistic regression Odds ratios (OR) Relative risk ratios (RRR) | Proportion of optimal birth weight (POBW), gestational age (GA), sex, socioeconomic status (SES), birth year, maternal age, group, maternal medical conditions, smoking. | Children born to mothers from foreign-born low-income countries had an increased relative risk of autism spectrum disorder (ASD) with intellectual disability (ID), and children born to foreign-born mothers from upper-middle-income countries had an increased risk of cerebral palsy (CP) with ID. | Indigenous mothers: risk ID [aRRR 1.75 (1.60–1.92)] and CP with ID [aRRR 1.74 (1.25–2.41)]. Foreign-born low-income mothers: risk ASD with ID [aRRR 2.34 (0.96–5.70)]. Upper-middle- income countries: risk ASD with ID [aRRR, 1.12 (0.70–1.78)]. | Med | |
| Camp et al. (83) United States | Multivariate regression Relative Risk (RR) Chi-squares Population attributable risks (PAR) | Race (Black and White | Low SES of the family accounted for 44–50% of ID and a low level of maternal education accounted for 20%. | Risk of ID for Black Children: Low SES RR 2.54, PAR 49.95% OR Low maternal education RR 2.34, PAR 20.06%. Risk of ID for White Children: Low SES RR 3.36, PAR 43.94% Low maternal education RR 3.00, PAR 19.65%. | Med | |
| Chapman et al. (82) Florida, United States | Risk ratio, Population attributable fraction | None | Low maternal education increased risk for all ID. Older maternal age increased risk of all ID, but for individuals with mild ID, this age effect was only seen in the lowest education group. | Mothers with <12 years education vs. some post-secondary education: RRR 7.00 (6.40–7.70). | Med | |
| Decoufle et al. (78) Atlanta, United States | Multiple logistic regression Exposure odds ratios | Child race, maternal education level and birth order. | General pattern of lower than expected risk of ID in children of white collar workers and higher risk in children of blue collar workers. | Risk of any ID in children of blue-collar workers in the textile products or apparel manufacturing industries: aOR 10.30 (1.30–81.70). | High | |
| Decoufle and Boyle (79) Atlanta, United States | Multiple logistic regression Odds ratios | Maternal race, age at delivery, child gender birth weight (BW), birth order, family SES. | Lower maternal education, race-education interaction. Maternal educational level strongly and inversely related to idiopathic ID. | Child risk of any ID for Black mothers with <10 years education: aOR 2.90 (1.60–5.40); White mothers: aOR 9.10 (3.90–21.30). Maternal education overall aOR 0.71 (0.65–0.79) Maternal education (White and Black combined) < 8 years aOR 6.10 (2.20–16.60) vs. > 16 years education aOR 0.20 (0.10–0.50) | High | |
| Drews et al. (75) United States | Logistic regression Odds ratios | Models included all 6 study variables. | Boys, children with two or more older siblings, Black children, and children whose mothers had not completed high school were more likely to have both mild and severe ID than were girls, firstborn children, White children, and children with college-educated mothers. Older maternal age was associated with an increased prevalence of severe ID; low SES with mild ID. ID with no biomedical cause was associated with high birth order and low maternal education. Older maternal age was associated with an increased prevalence of ID with defined etiologies. Results held when dividing by IQ group (i.e., mild 50–70, severe < 50). | Mild ID (n = 330) Male sex: aOR 1.60 (1.20–2.20) Birth order > 3rd: aOR 1.60 (1.10–2.50) Black race: aOR 1.80 (1.30–2.60) Low SES: aOR 1.60 (1.00–2.50) Maternal education <12 years: aOR 4.10 (2.40–6.90) Maternal education = 12 years: aOR 1.80 (1.10–2.90) Severe ID (n = 128) Male sex: aOR 1.70 (1.10–2.50) Maternal age > 30 years: aOR 1.80 (1.10–3.10) Isolated ID (n = 316) Male sex: aOR 1.70 (1.20–2.30) Birth order > 3rd: aOR 1.90 (1.20–2.90) Black race: aOR 2.30 (1.60–3.30) Maternal education < 12 years: aOR 4.90 (2.80–8.50) Maternal education = 12 years: aOR 2.10 (1.20–3.40) | Med | |
| Emerson (74) England | Multilevel logistic regression Odds ratio | Ethnicity, household deprivation and area deprivation. | Lower household SES was associated with increased likelihood of mild-moderate and severe ID but lesser likelihood for profound multiple ID. Higher rates of mild-moderate ID were found in Gypsy/Romany and Traveler children of Irish heritage and of severe iD in children of Pakistani heritage, Minority ethnic status was otherwise associated with lesser likelihood of ID. | Area deprivation codes d1-d9 aORs significant for mild ID (aOR 3.48 (3.38–3.57) at most deprived area) and severe ID (aOR 1.45 (1.36–1.55 at worst area). | Med | |
| Heikura et al. (69) Northern Finland | Multivariate logistic regression Odds ratios | Maternal age, parity, body mass index, marital status, SES, and place of residence | Socioeconomic disadvantage and maternal multiparity had greatest impact on risk of ID. Pre-pregnancy maternal obesity became evident in the more recent cohort. | 1966 cohort, risk for mild ID: SES IV Level 4 Unskilled OR 1.80 (0.60–5.30); Multiparity (parity > 4) OR 1.30 (0.70–2.60). 1986 cohort: SES IV Level 4 Unskilled OR 2.70 (1.50–4.70), Multiparity (parity > 4) OR 3.30 (1.80–6.00). | For any ID, in 1966 cohort: Multiparity (parity > 4) aOR 1.50 (0.60–3.60); 1986 cohort: Multiparity (parity=4) aOR 2.40 (1.20–4.90) Pre-pregnancy body mass index (BMI) aOR 2.80 (1.50–5.30) For any ID: 1966 cohort: SES IV Level 4 unskilled; aOR 1.70 (0.50–5.80). 1986 cohort: SES IV Level 4 Unskilled, aOR 2.20 (1.20–3.90) For mild ID, being a farmer in 1966 cohort aOR 3.50 (1.10–11.40); and in 1986 cohort aOR 3.70 (1.40–9.80). For severe ID, living in a remote area in 1966 cohort aOR 1.90 (1.00–3.60), level IV unskilled worker for 1986 cohort aOR 3.10 (1.40–6.70). | High |
| Leonard et al. (60) Western Australia, Australia | Logistic regression Odds ratios | Birth year | Aboriginal mothers, teenagers and single mothers at increased risk of having a child with mild or moderate ID. Children of mothers in the most socioeconomically disadvantaged 10% had more than five times the risk of mild and moderate ID than those in the least disadvantaged 10%. Fourth or later born children were also at increased risk. | Mild-moderate ID: Aboriginal mothers OR 2.83 (2.52–3.18), teenagers OR 2.09 (1.82–2.40) and single mothers 2.18 (1.97–2.42). Most socioeconomic disadvantaged 10%: OR 5.61 (4.40–7.12), 6th or later-born infants OR 3.13 (2.50–3.91). | Med | |
| Leonard et al. (62) Western Australia, Australia | Multinomial logistic regression Odds ratios | Birth year, POBW, sex, birth order, maternal ethnicity, country of birth, height, marital status, maternal and paternal age, SES, geographical remoteness. | Gradient effect where risk factors for mild-moderate ID and ASD without ID were at opposite extremes. Mild-moderate ID was associated with younger mothers <20 years, paternal age >39 years, Australian-born and Aboriginal mothers, increasing birth order and increasing social disadvantage. Mothers of infants residing in regional or remote areas had consistently lower risk of ID. | Mild-moderate ID Maternal age <20 years, OR 2.41 (2.17–2.69); Paternal age <20 years, 2.38 (1.98–2.87); Aboriginal OR 1.79 (1.15–2.78); Low SES OR 2.56 (2.27–2.97; 4th. or later born OR 2.06 (1.89–2.25). | Mild-moderate ID Maternal age < 20 years, aOR 1.88 (1.57–2.25); paternal age >39 years, aOR 1.59 (1.36–1.86); Aboriginal aOR 1.60 (1.41–1.82; low SES aOR 2.56 (2.27–2.97); 4th or later born, aOR 2.55 (2.27–2.87). | Med |
| Oswald et al. (43) Columbia, United States | Logistic regression Odds ratios | District weights. | ID more prevalent in males but relationships with gender vary by racial group. | ID by Gender/Ethnicity Groups: Male American Indian: OR 1.66 Male Black: OR 3.26 Male White: OR 1.36 Female Black: OR 2.02 | ID by Gender/Ethnicity Groups: Male American Indian: aOR 1.60 Male Black: aOR 4.03 Male Hispanic: aOR 1.35 Female Black: aOR 2.58 | Med |
| Rantakallio (42) Northern Finland | Pearson chi square Proportions | None | Mild and severe ID associated with lower class and farmers. | Idiopathic ID Social class I & II Inc. mild ID 0.4 per 1,000 Inc. moderate ID 0.4/1,000 Social class IV Inc. mild ID 2.5/1,000 Inc. moderate ID 1.9/1,000 | Low | |
| Stromme and Magnus (38) Norway | Logistic regression Odds ratios | None | Lower SES was associated with mild ID. Severe ID associated with higher SES. | Children with parents in Social Classes IV to V had increased risk of mild-moderate ID compared with severe ID OR 6.30 (0.70–23.50) for IV OR 5.00 (1.50–15.60) for V | Med | |
| Williams and Decoufle (33) Atlanta, United States | Logistic regression Odds ratios | Child gender, BW and birth order, maternal age, education, race, father presence. | Increased risk of biomedical ID was seen among Black children of mothers aged >30 years not attributable to Down syndrome | Biomedical ID: Maternal age <20 years: OR 2.82 (1.51–5.26) Idiopathic ID Maternal age <20 years: OR 1.39 (0.76–2.47) | Biomedical ID Maternal age <20 years: aOR 2.20 (1.06–4.59) Idiopathic ID Maternal age <20 years: aOR 0.85 (0.40–1.83) | High |
| Yeargin-Allsopp et al. (32) Atlanta, United States | Logistic regression Odds ratios | Sex, maternal age, birth order, maternal education, SES. | Black race, but not for children diagnosed before 6 years of age. | Risk of mild ID in Black compared with White children: OR 2.60 (2.00, 3.50) | Risk of mild ID in Black compared with White children: OR 1.80 (1.30–2.60) | High |
| Zhen et al. (31) United States | Bayesian local likelihood cluster modeling RRs Chi-square p-values | None | Using spatial analysis for cluster detection according to mother’s geocode during her pregnancy a of ID was higher than surrounding areas. The descriptive analysis of the study population showed that in general children with ID were more likely to be male and had mothers who were older than 34 years at the time of birth as well as being African American, preterm and of low birth weight compared to children without ID. | Rate per 100 for ID cases in study area in South Carolina (n), ID cases in cluster area (n) Month 1: 4.3 (1287), 5.9 (185). Month 2: 4.1 (2213), 7.0 (273) Month 3: 4.0 (2813), 8.0 (337) Month 4: 3.9 (3169), 8.5 (366) Month 5: 3.9 (3331), 7.7 (388) Month 6: 3.8 (3386), 7.6 (395) Month 7: 3.6 (3225), 7.1 (408) Month 8: 3.4 (2887), 7.0 (370) Month 9: 3.2 (2091), 5.3 (281) Month 10: 3.8 (523), 8.9 (79) | Low | |
| Zheng et al. (30) China | Logistic regression Odds ratios | Age, gender, parent disability | Male sex, mother’s age, Illiteracy, rural residence, parental ID were risk factors for both mild and severe ID as. Family income lowest compared with highest showed a monotonic increase in the likelihood of ID. | Mild ID: Male OR 1.22 (1 04–1.42) Maternal age >34 vs. 20–29 OR 1.77 (1.35–2.31) Mother illiterate vs. Senior High School or higher OR 4.53 (3.63–5.66) Rural vs. Urban residence OR 2.21 (1.86–2.63) Parental ID vs. Non-ID OR 14.67 (10.74–20.05) Severe ID: Male OR 1.13 (0.96–1.32). Maternal age >34 vs. 20–29 OR 1.53 (1.14–2.05) Mother illiterate vs. Senior High School or higher: OR 5.67 (4.51–7.13) Rural vs. Urban residence OR 1.73 (1.44–2.09) Parental ID vs. Non-ID OR 13.81 (9.87–19.30) | Mild ID: Male vs. female aOR 1.23 (1 06–1.44) Maternal age > 34 vs. 20–29 aOR 1.47 (1.15–1.88) Mother illiterate vs. Senior High School or higher aOR 1.74 (1.21–2.52) Rural vs. Urban residence aOR 1.93 (1.55–2.41) Parental ID vs. Non-ID aOR 8.81 (6.45–12.06) Severe ID: Male vs. female aOR 1.14 (1.00–1.73). Maternal age > 34 vs. 20–29 aOR 1.32 (1.14–2.05) Mother illiterate vs. Senior High School or higher aOR 2.45 (1.65–3.63) Rural vs. Urban residence aOR 1.35 (1.09–1.67) Parental ID vs. Non-ID aOR 7.10 (5.09–9.91) | Med |
| Antenatal and perinatal | ||||||
| Bilder et al. (85) Utah, United States | Logistical regression Multiple logistic regression. | Maternal and paternal age, education, GA and parity | Poly/oligohydramnios, advanced paternal/maternal age, prematurity, fetal distress, premature rupture of membranes, primary/repeat cesarean sections, low birth weight, assisted ventilation greater than 30 min, small-for-gestational age, low Apgar scores, and congenital infection, associated with increased odds of mild to severe ID. | Risk of mild and severe ID Maternal age > 34 OR 1.82 (1.20–2.76) Paternal age > 34 OR 2.00 (1.38–2.90) Uterine bleeding OR 2.63 (1.07–6.4) Poly/oligohydramnios OR 5.45 (2.51–11.84). Maternal education > 13 years OR 0.67 (0.47–0.96) | Risk of mild and severe ID GA < 37 weeks aOR 3.07 (1.98–4.76) Primary cesarean section aOR 3.95 (2.48–6.30) Premature rupture of membranes aOR 2.48 (1.20–5.13) Fetal distress aOR 3.05 (1.75–5.30) Low birth weight < 2500 gm aOR 3.43 (2.21–5.33) Low 5-min Apgar score < 7 aOR 8.11 (4.02–16.36) Assisted ventilation > 30 min aOR 5.35 (2.61–10.98) | High |
| Chen et al. (80) Sweden | Cox regression models. Hazard ratios (HR) | Maternal age, education parity, country of birth, smoking, height, weight, labor onset, mode of delivery, sex, birth period, | Among appropriate-for-gestational-age (AGA) children born at term or post-term, lower birth weight percentiles within the normal range are associated with increased risk of ID, regardless of GA. | Compared with the reference group (40th–59th birthweight percentiles): 10th–24th percentiles aHR 1.43 (1.22–1.67) (population analysis) aHR 1.52 (1.00–2.31) (sibling analysis) 25th–39th percentiles: aHR 1.28 (1.10–1.50) in (population analysis) aHR 1.44 (1.00–2.09) (sibling analysis) | High | |
| Chen et al. (81) United States | Logistic regression. Odds ratios. | Model 1: maternal age, race, SES, education, pre-pregnancy BMI, smoking, parity, sex, GA. | Placental inflammation was associated with risk of ID at 7 years of age. No indirect effects of shorter GA. | Risk of ID at 7 years of age: aOR 1.20 (0.97–1.48) | Low | |
| Croen et al. (13) California, United States | Poisson regression. Risk ratios. Adjusted RRs. | Gender, weight, plurality, parity, maternal age, race/ethnicity, education, birthplace. | For both children with mild and severe ID, risk was increased among those who were male, low birth weight, born to women of Black race, older age at delivery, and lower level of education. Increased risk for mild ID was found for multiple births, second or later-born children, and children with mothers born outside of California. Children born to Hispanic mothers had increased risk of severe ID, as had children born to Asian mothers who also had decreased risk of mild ID. | Mild ID: Male OR 1.90 (1.80–2.00) Low birth weight (LBW) OR 4.90 (4.60–5.20) Black OR 2.10 (1.90–2.30) Maternal age > 35 OR 1.00 (0.90–1.20) High education OR 0.30 (0.30–0.50) Multiple births OR 3.00 (2.70–3.30) Later-born OR 1.60 (1.50–1.70) Asian OR 0.60 (0.50–0.70) Severe ID: Hispanic OR 1.40 (1.10–1.80) | Mild ID: Male aOR 1.90 (0.80–2.00) LBW aOR 4.30 (4.00–4.60) Black aOR 1.50 (1.40–1.70) Maternal age > 35 aOR 1.40 (1.20–1.60) High education aOR 0.40 (0.30–0.40) Multiple births aOR 1.20 (1.10–1.40) Later-born aOR 1.30 (1.20–1.50) Asian aOR 0.70 (0.60–0.80) Severe ID: Hispanic aOR 1.20 (1.00–1.40) Asian aOR 1.30 (1.00–1.70) | High |
| Heuvelman et al. (68) Sweden | Multivariate regression, Generalized estimating equations (GEE) Odds ratios | Sex, parity, maternal age, gestational diabetes (GDM), hyper-tension, pre-eclampsia, parental country of birth, psychiatric history, family income, family size, parental education. | Risk of ID was greatest in those born extremely early, lessening with advancing GA but increasing post term. Risk of ID was greatest among those showing evidence of fetal growth restriction especially before or after term. | Risks of ID At 24 weeks GA aOR 14.54 (11.46–18.44]. At 32 weeks GA aOR 3.59 (3.22–4.01]. At 37 weeks GA aOR 1.50 (1.38–1.63) At 39 weeks GA aOR 1.10 (1.04–1.17) At 42 weeks aOR 1.16 (1.08–1.25) At 44 weeks aOR 1.71 (1.34–2.18) | High | |
| Jones et al. (65) California, United States | Multivariate logistic regression. Odds ratios | Maternal age, ethnicity, birth country, weight, sex, birth year and month | In comparison to general population, higher mid-gestational maternal serum levels of GM-CSF, IL-1α, IL-6, and IFN-γ associated with an increased risk of ASD + ID. The principal association in the DD group relative to GP controls was a decrease in risk with elevated levels of IL-8 and MCP-1. | The only significant associations in the ID group relative to GP controls was a decrease in risk with elevated levels of IL-8 and MCP-1. | ID vs. GP (reduced) IL-1B: aOR 0.91 (0.83–1.00) IL-8: aOR 0.81 (0.70–0.93) MCP: aOR 0.76 (0.59–0.98) | High |
| Langridge et al. (64) Western Australia, Australia | Logistic regression Odds ratios | Birth year, maternal pregnancy, labor and delivery factors, neonatal outcomes, sociodemographics. | Mild-moderate ID was associated with pregnancy hypertension, asthma, urinary tract infection, some types of antepartum hemorrhage (APH), any type of preterm birth, elective cesarean-sections, breech presentation, poor fetal growth and need for resuscitation at birth. Severe ID was associated with poor fetal growth and need for resuscitation, and any labor or delivery complication. | Mild-moderate ID: Asthma aOR 1.28 (1.12–1.47) Urinary tract infections aOR 1.36 (1.19–1.55) Other APH aOR 1.41 (1.19–1.68) Pregnancy hypertension aOR 1.39 (1.25–1.54), Spontaneous/pre-labor rupture of membranes aOR 1.63 (1.42–1.87) Medically indicated Preterm birth aOR 1.25 (1.04–1.49) Elective CS aOR 1.17 (1.03–1.32) Breech presentation aOR 1.33 (1.13–1.56) POBW < 75%, mild ID aOR 2.33 (1.97–2.75) and severe ID aOR 3.77 (2.15–6.61) Resuscitation at birth, mild ID aOR 1.41 (1.31–1.52) and severe ID aOR 2.18 (1.61–2.94). Male sex, mild ID aOR 1.61 (1.50–1.73) and severe ID aOR 1.67 (1.24–2.24). | Med | |
| Leonard et al. (61) Western Australia, Australia | Multivariate logistic regression Odds ratios | Marital status, maternal country of birth, health insurance status, paternal occupation, geographic remoteness, birth year, and SES | Inappropriate intrauterine growth, less than or greater than optimal birth weight, increased risk of ID. Effects were similar among Caucasian and Aboriginal children. | Caucasian: Mild-moderate ID: Severe growth restriction preterm (<37 weeks) aOR 1.71 (1.06–2.77). Severe ID: Term births aOR 4.79 (2.59–8.83) Growth restriction aOR 3.20 (1.30–7.90) Poor head growth aOR 3.60 (1.40–9.00) | Med | |
| Louhiala (58) Central and Southern Finland | Logistic regression, Multi-variate analyses. Odds ratios | Adjustment was made in multivariate analysis but covariates unknown | Increased risk of any ID with multiparity, low GA, multiple pregnancy, low social class, male sex, hyperbilirubinemia, hypoglycemia, SGA. Smoking, hypertension, preterm birth (PTB), and maternal advanced age did not increase adjusted odds of ID. | Risk of any ID GA < 33 weeks: OR 17.10 (2.30–130.40) 15 min Apgar score < 7: OR 14.40 (1.80–113.80) Maternal epilepsy: OR 12.00 (0.70–217.60) 5 min Apgar score < 7: OR 7.90 (2.70–22.90) Hypoglycemia: OR 6.50 (2.70–15.60) Multiple birth: OR 6.00 (1.70–20.60) BW < 1500 g: OR 5.60 (1.20–25.70) SGA: OR 4.50 (2.60–8.90) BW < 2500 g: OR 3.40 (1.90–6.10) GA < 35 weeks: OR 2.80 (1.30–5.80) Hyperbilirubinemia: OR 2.70 (1.20–4.10) Breech presentation: OR 2.70 (1.20–4.10) Male sex: OR 1.60 (1.20–2.10) Lowest social class: OR 1.40 (1.90–1.90) Multiparity (>2): OR 1.30 (1.00–1.80) | Risk of any ID GA < 35 weeks: aOR 0.50 (0.20–1.40) Hypoglycemia: aOR 3.60 (1.40–3.70) 1-min Apgar score < 7: aOR 4.80 (2.30–10.10) SGA: aOR 4.40 (2.10–9.30) Multiple birth: aOR 5.60 (1.40–22.30) Hyperbilirubinemia: aOR 3.00 (1.20–7.60) Male sex: aOR 1.50 (1.10–2.10) Lowest social class: aOR 1.70 (1.20–2.40) Multiparity (<2): aOR 1.50 (1.00–2.10) | Low |
| Luu et al. (57) Providence, Portland and New Haven United States | Linear, logistic regression. Between group means | Sex, maternal age, maternal education, minority status, household structure. | Preterm children obtained scores 6-14 points lower than term controls on all psychometric tests. Severe neonatal brain injury was the strongest predictor of poor intelligence. Antenatal steroids, higher maternal education, and 2-parent family were associated with better cognition. | IQ adj. mean diff for all children preterm vs. term: -13.70 (-17.10, -10.30) (p < 0.005) IQ adj. mean diff preterm vs. term excluding brain injury children: -11.70 (-14.70, -8.70) (p < 0.005) | Low | |
| McDermott et al. (50) California, United States | Multiple logistic regression Risk ratios | Maternal age, race, education, smoking, alcohol use, pregnancy conditions, congenital infant anomalies, resuscitation at birth. | Strong association between LBW and mild ID for both Blacks and non-Blacks at age 5. Association only held for Blacks in the cohort of children aged 9–11. | Risk of ID in LBW (<2500 g) infants at age 5 years: All aOR 3.40 (2.10–5.40) Black aOR 2.70 (1.40–5.40) Non-Black aOR 3.90 (2.00–7.50) Risk of ID in LBW (<2500 g) infants at age 9–11 years All aOR 1.20 (0.70–2.00) Black aOR 4.10 (1.40–12.00) Non-Black aOR 0.80 (0.40–1.60) | Med | |
| Mervis et al. (47) Atlanta, United States | Logistic regression Odds ratios | Child sex, birth order, GA, maternal age, race, education | Risk of ID was increased for LBW (<2500 g) infants, higher for very low birth weight (VLBW) (<1500 g) and for severe ID. Preterm infants with normal birthweight also had increased risk. | Risk of mild ID: <2500 g aOR 2.40 (1.60–3.80) <1500 aOR 11.70 (2.50–54.70) Risk of severe ID: <2500 g aOR 4.40 (2.60–7.50) <1500 g aOR 29.40 (5.80–148.20) | High | |
| Schieve et al. (40) United States | Multivariate logistic regression Relative risks and standardized morbidity ratios | Child sex, birth order, maternal race-ethnicity, education, age, marital status, and smoking during pregnancy | Children with any ID had higher rates of PTB, LBW, SGA, and low Apgar score than expected based on the US birth cohort, with a higher percentage of mothers who smoked during pregnancy. | Relative risks for any ID PTB SMR aRR 2.10 (1.90–2.20) VPTB aRR 5.40 (4.80–5.90) LBW aRR 3.20 (3.00–3.40) VLBW aRR 7.80 (7.00–8.70) Term aRR LBW 2.40 (2.10–2.70) SGA aRR 1.90 (1.80–2.00) VSGA aRR 2.30 (2.10–2.50) Term SGA aRR 1.60 (1.50–1.70) Low Apgar aRR 8.50 (7.60–9.50) HBW aRR 0.60 (0.50–0.70) Large for GA aRR 0.60 (0.50–0.70) | High | |
| Van Naarden et al. (35) United States | Observed vs. expected ratios | Maternal age | The proportions of multiple births in children with ID born in three US states in 1994 were moderately higher than expected based on US population data. | Observed/expected multiple birth 1.34 [0.95–1.73]). | High | |
| Maternal physical health | ||||||
| Blotiere et al. (84) France | Number of events, crude event rates and crude incidence rate ratios (IRRs) | Maternal age, medication, co-morbidities, history of mental or behavioral disorders, sex, GA, BW. | Association between maternal use of valproic acid (VPA) during pregnancy and the risk of ID in offspring, with a dose-response relationship. No increased risk of any of the neurodevelopmental outcomes for prenatal exposure to any other antiepileptic drugs (AEDs) studied. | Compared with prenatal exposure to lamotrigine VPA had increased risk of isolated ID aHR = 3.10 (1.50–6.20) | Med | |
| Drews et al. (76) Atlanta, United States | Logistic regression Odds ratios | Sex, maternal age, race, education, SES, parity, and alcohol use. | Smoking during pregnancy was associated with an increase in the prevalence of ID and the risk increased if at least one pack was smoked a day. | Risk of idiopathic ID with smoking at any time during pregnancy: OR 1.74 (1.21–2.50) Risk of idiopathic ID with smoking during 2nd trimester: OR 1.75 (1.20–2.55) | Smoking at any time during pregnancy: aOR 1.44 (0.92–2.27) Smoking during 2nd trimester: aOR 1.53 (0.80–2.46) | High |
| Griffith et al. (71) South Carolina, United States | Multiple logistic regression Odds ratios | Maternal age, race, education, birth year and sex | Pre-eclampsia was associated with an increase in the odds of ID. LBW was a mediator of the relationship | Risk of any ID in mothers with pre-eclampsia (PET): OR 1.58 (1.33–1.87) | Risk of any ID in mothers with PET: aOR 1.38 (1.16–1.64) Risk of any ID in mothers with PET: aOR 1.28 (1.07–1.52) including LBW in the model | High |
| Leonard et al. (63) Western Australia, Australia | Logistic regression Odds ratios | Infant sex and birth order, maternal ethnicity, age group, marital status, height, country of birth, health insurance status, paternal occupation. Remoteness, aggregated SES measure. | Increased risk for mild to moderate ID in children whose mothers had renal or urinary conditions, asthma, and diabetes (ORs, 1.23–1.65) and a threefold increased risk for women with epilepsy. Fivefold risk for having a child with severe for mothers with anemia. | Asthma in mothers: mild to moderate ID (OR, 1.52; CI, 1.26–1.83). Diabetes: mild to moderate ID (OR, 1.69; CI, 1.26–2.27). Undiagnosed cardiac murmurs: (mild to moderate ID, OR, 0.99; CI, 0.61–1.60; severe ID, OR, 2.04; CI, 0.65–6.39) Hypertension: mild to moderate (OR, 0.99; CI, 0.58–1.68) or severe ID (OR, 2.48; CI, 0.79–7.77). Renal or urinary condition mild to moderate ID (OR, 2.09; CI, 1.39–3.14), but not severe ID (OR, 1.01; CI, 0.14–7.19). Epilepsy: mild-moderate ID (OR, 3.53; CI, 2.56–4.84). Anemia: severe ID (OR, 5.26; CI, 2.16–12.80). | Stepwise adjusted logistic regression model with mild-moderate ID as outcome: maternal epilepsy (OR, 3.01; CI, 2.10–4.33), maternal renal or urinary condition (OR, 1.65; CI, 1.06–2.56), and maternal asthma (OR, 1.25; CI, 1.02–1.54), maternal diabetes (OR, 1.38; CI, 0.99–1.91; p = 0.06). For severe ID, maternal anemia: (OR, 4.93; CI, 2.00–12.09). | Med |
| Li et al. (59) Boston, United States | Cox regression Hazard ratios | Child birth year, sex, maternal age, parity, smoking and PTB. | Obesity, pre-gestational diabetes (PGD), obesity with PGD and obesity with gestational diabetes (GD) were all associated with an increased risk of ID. | Risk of any ID Obesity: aHR 1.64 (1.09–2.45) PGD: aHR 2.26 (1.25–4.09) Obese & PGD: aHR 3.63 (1.73–7.61) Obese & GD: aHR 2.31 (1.00–5.36) | High | |
| Mann et al. (55) South Carolina, United States | Cox proportional hazards Models Hazard ratios | Maternal age, race, education level, alcohol use, smoking, sex. | Trichomoniasis during pregnancy associated with ID. Second trimester trichomoniasis associated with a threefold increase in odds of mild or severe ID in the public school system | Trichomoniasis in pregnancy Risk of ID: aHR 1.28 (1.12–1.46) Risk of severe ID: aHR 1.83 (1.34–2.40) | High | |
| Mann et al. (54) South Carolina, United States | Multiple logistic regression & GEE Odds ratios | Maternal age, race, ethnicity, education, smoking, sexually transmitted disease, hyper-tension, diabetes, epilepsy, GA, BW. | The risk of ID was greater in children of women with pre-pregnancy obesity; greatest for women with morbid obesity for ID of any severity. Gestational weight change (gain or loss) was not associated with odds of ID. | Risk of any ID. Pre-pregnancy morbid obesity: aOR 1.52 (1.30–1.77) Risk of severe ID. Pre-pregnancy morbid obesity: aOR 1.83 (1.31–2.56) | Med | |
| Mann et al. (53) South Carolina, United States | Logistic regression Odds ratios GEE | Maternal race, education, age, smoking, hyper-tension, sex, GA, BW. | Maternal diabetes associated with an increased risk of ID in offspring. | Risk of ID Any diabetes: aOR 1.09 (1.01–1.19) Chronic diabetes: aOR 1.31 (0.83–2.06) GDM: aOR 1.08 (0.98–1.19) Diabetes uncertain timing: aOR 1.12 (0.95–1.31) | Med | |
| McDermott et al. (51) South Carolina United States | Logistic regression Relative risk | Maternal age, race, alcohol use, infant GA. | Mothers experiencing a UTI during pregnancy had a slightly increased risk of a child with ID. The risk was increased if the mother did not receive antibiotic treatment. | Risk of ID following UTI in pregnancy depending on antibiotic treatment adjusted only for GA. Unfilled script vs. No UTI controls: RR 1.31 (1.12–1.54) Unfilled script vs. filled script for UTI: RR 1.22 (1.02–1.46) | Risk of ID following UTI in pregnancy depending on antibiotic treatment with full adjustment. Unfilled script vs. No UTI controls: aRR 1.31 (1.12–1.54) Unfilled script vs filled script for UTI: aRR 1.22 (1.02–1.47) | Med |
| McDermott et al. (49) South Carolina United States | Logistic regression and survival analysis Relative risk | Maternal education, race. Infant sex, GA, BW. | Risk of child ID was increased in women with a urinary tract infection (UTI) during pregnancy | Risk of ID following UTI in pregnancy Overall: aRR 1.16 (1.00–1.29) Trimester 3: aRR 1.40 (1.01–1.95) | High | |
| Salonen and Heinonen (41) Finland | Logistic regression Risk ratios | Maternal age, parity, smoking, mode of delivery, birth defects in siblings. | A strong association was found between pregnancy hypertension and risk of ID in the child | Pregnancy hypertension: risk of any ID, RR 6.00 (1.30–27.00) | Pregnancy hypertension: risk of any ID, aRR 6.10 (1.30–28.90) | Low |
| Takei et al. (37) England and Wales | Generalized linear model, Poisson regression. Risk ratios | N/A. | Increased death rates from influenza were significantly associated with an increase in births of ID individuals 6 months later. | For every 1,000 female deaths from influenza there was a 17% increase in births of individuals with ID 6 months later. RR 1.17 (3.70–31.70) | Low | |
| Tomson et al. (36) Sweden | Multivariate logistic and Cox models with covariates. Adjusted ORs and HRs | Infant age, sex, GA, BW for GA, parity, maternal education, smoking during pregnancy, cohabitation of parents, parental age, parental psychiatric history. | No evidence for an association between paternal exposure to AEDs during the period of spermatogenesis and increased risk for ID in the offspring. But risk of ID increased in children of fathers with epilepsy, this was not related to the AED medication. High risk of ID amongst mothers with epilepsy, and higher risk for those exposed to AEDs during pregnancy, especially VPA. | All for risk of any ID (mild to severe): Fathers with epilepsy, risk of ID aHR 1.80 (1.20–2.90) vs. without epilepsy. Mothers with epilepsy, risk of ID aHR 2.40 (1.80–3.10) vs. mothers without epilepsy. Mothers with epilepsy who used AED during pregnancy, risk of ID for VPA aHR 9.60 (3.50–26.20), for carbamazepine aHR 4.70 (1.80–12.50), for “other” AED aHR 4.70 (1.40–16.40), for lamotrigine aHR 2.50 (0.90–7.20), vs. unexposed to any AED. | High | |
| Maternal mental health | ||||||
| Di Prinzio et al. (77) Western Australia, Australia | Multiple logistic regression Odds ratios | Sex, birth year, parental age, race, place of birth, birth order, marital status, SES, residential remoteness, father’s psychiatric status. | Risks of ID was increased among children of mothers with severe mental illness compared with children of unaffected mothers. | Risk of ID for children of mothers with any severe mental illness OR 2.70 (2.40–3.00) Risk of ID for children of mothers with schizophrenia OR 3.80 (3.00–4.90) Risk of genetic ID for children of mothers with any severe mental illness OR 2.00 (1.54–2.70) Risk of genetic ID for children of mothers with schizophrenia OR 2.40 (1.20–5.10) | Risk of ID for children of mothers with any severe mental illness aOR 1.70 (1.50–1.90) Risk of ID for children of mothers with schizophrenia aOR 1.70 (1.30–2.30) Risk of genetic ID for children of mothers with any severe mental illness aOR 1.60 (1.20–2.20) Risk of genetic ID for children of mothers with schizophrenia aOR 1.60 (0.70–3.60) | High |
| Fairthorne et al. (72) Western Australia, Australia | Multinomial logistic regression, binary logistic models for composite case groups. Crude and adjusted ORs. | Maternal age, parity and the index birth year group | Compared to mothers with no previous psychiatric contact, those with any psychiatric contact were more than twice as likely to have a child with ID. | Any previous psychiatric contact (compared to mothers without): ID unknown cause aOR 2.14 (1.90–2.40), ID known cause (not DS) aOR 2.13 (1.60–2.80), ASD with ID aOR 1.77 (1.30–2.40). Schizophrenic disorders: any ID aOR 2.53 (1.40–4.50); ID of unknown cause aOR 2.20 (1.10–4.40). Affective disorders: ASD with ID aOR 2.23 (1.50–3.20), ID unknown cause aOR 1.89 (1.60–2.20); ID known cause (not DS) aOR 1.71 (1.10–2.70). | Med | |
| Morgan et al. (46) Western Australia, Australia | Logistic regression Odds ratios | Child sex, indigenous status and birth order | Children of mothers with schizophrenia, bipolar disorder and unipolar depression were at risk of having offspring with ID certain neonatal complications were independent predictors. | Risk of ID in mothers with Schizophrenia: OR 3.20 (1.80–5.70) Bipolar disorder: OR 3.10 (1.90–4.90) Unipolar depression: OR 2.90 (1.80–4.70) | Risk of ID in mothers with: Schizophrenia: aOR 2.20 (1.20–4.30) Bipolar disorder: aOR 2.60 (1.50–4.40) Unipolar depression: aOR 2.70 (1.60–4.50) | Low |
| O’Leary et al. (45) Western Australia, Australia | Logistic regression GEE Odds ratios Population attributable fractions | Maternal age and birth year, marital status, parity, maternal mental health, maternal illicit drug use, remoteness, SES. | There was a threefold increase in the adjusted odds of ID in children of mothers with an alcohol-related diagnosis recorded during pregnancy. At least 3.8% of cases of ID could be avoided by preventing maternal alcohol use disorder. | Risk of any ID: Non-Aboriginal: Any alcohol diagnosis OR 1.81 (1.53–2.14); During pregnancy OR 3.52 (1.96–6.34) Aboriginal: Any alcohol diagnosis: OR 1.66 (14.2–1.96); During pregnancy: OR 3.12 (2.13–4.56) | Risk of any ID: Non-Aboriginal: Any alcohol diagnosis: aOR 1.44 (1.18–1.75); During pregnancy: aOR 2.89 (1.62–5.18) Aboriginal: Any alcohol diagnosis: aOR 1.66 (14.2–1.96); During pregnancy: aOR 3.12 (2.13–4.56) | Med |
| Wang et al. (34) South Carolina, United States | Logistic regression GEE models Odds ratios | BW, birth year, maternal age, race, education, BMI, smoking, alcohol use, drug use, nutrition program participation during pregnancy, maternal ID, asthma, anemia, hyper-tension, myalgia, genito-urinary infection, APH, chorioamnionitis, early or threatened labor. | Women with bipolar disorder or major depression had increased odds of child having ID. Male children higher for bipolar, female child higher for major depression. | Mother had major depression: For males aOR 1.34 (1.14–1.59) PAR 2.17% For females aOR 1.59 (1.30–1.95) PAR 4.70%, Maternal bipolar disorder: For males aOR 1.95 (1.53–2.48) PAR 2.85% For females aOR 1.63 (1.20–2.22) PAR 2.05% | Low | |
| Environmental | ||||||
| Emerson et al. (73) United Kingdom | Non-parametric tests; exposure/prevalence rate ratios. | Complex samples module and sample weights to adjust for the initial sampling design and biases in recruitment and retention at each wave. | British children with ID are more likely than their peers to live in localities with high rates of outdoor air pollution. Averaging across ages, children with IDs were 30–33% more likely to be exposed to high rates of diesel particulate matter, nitrogen dioxide, and carbon monoxide, and 17% more likely to be exposed to high rates of sulfur dioxide. | Exposure rate ratios: 9 months (ID = 552, other = 18,000) PM10: 1.31 (1.14–1.50) NO2: 1.25 (1.08–1.44) SO2: 1.15 (0.93–1.39) CO: 1.23 (1.07–1.41) 3 years (ID = 525, other = 14,371) PM10: 1.35 (1.16–1.56) NO2: 1.31 (1.13–1.51) SO2: 1.16 (0.94–1.43) CO: 1.33 (1.16–1.53) 5 years (ID = 528, other = 14,384) PM10: 1.38 (1.20–1.59) NO2: 1.33 (1.15–1.53) SO2: 1.18 (0.98–1.43) CO: 1.37 (1.19–1.58) | Low | |
| Mackay et al. (56) Scotland | Binary logistic regression. Odds ratios | Maternal age, SES, parity, PET, previous spontaneous abortion, GA, sex and gestation-specific BW percentiles, mode of delivery | Rates of ID highest among children conceived in the first quarter of the year (January–March) and lowest among those conceived in the third (July–September). | Association between calendar year quarter of conception and likelihood of ID (referent 3rd quarter) Quarter 1: OR 1.23 (1.18–1.29) Quarter 2: OR 1.18 (1.13–1.23) Quarter 4: OR 1.13 (1.09–1.18) | Association between calendar year quarter of conception and likelihood of ID (referent 3rd quarter) Quarter 1: aOR 1.23 (1.18–1.29) Quarter 2: aOR 1.17 (1.12–1.22) Quarter 4: aOR 1.13 (1.09–1.18) | High |
| McDermott et al. (48) South Carolina, United States | Logistic regression Generalized additive model Odds ratios | Sex, GA, BW, SGA, maternal age, race, parity, smoking, alcohol use. | The probability of ID increased for increasing concentrations of arsenic (As) and lead (Pb) in the soil. | Increased risk of ID for I unit in change in As soil concentration aOR 1.13 (1.05–1.22) Appropriate for GA aOR 1.15 (1.06–1.25) Increased risk of ID for I unit in change in Pb oil concentration aOR 1.002 (1.000–1.004) Appropriate for GA aOR 1.002 (1.002–1.004) | Med | |
| McDermott et al. (52) South Carolina, United States | Logistical models. Generalized additive models Odds ratios | Sex, GA, BW, SGA, maternal age, race, parity, smoking, alcohol use. | Positive association between increased soil mercury (Hg) and mild ID; and between soil As, barium, and lead levels and severe ID. Hg and As associated with all levels of ID. In the nine case areas of research, ID prevalence was 5.9%. In the control area the prevalence was 3.2%. | Crude odds ratios Arsenic: Mild ID vs. no ID: 1.00 (0.95–1.05) Severe ID vs. no ID: 1.07 (1.02–1.12) Any ID vs. no ID: 1.03 (1.00–1.07) Mercury: Mild ID vs. no ID: 1.71 (1.34–2.19) Severe ID vs. no ID: 0.91 (0.61–1.35) Any ID vs. no ID: 1.35 (1.09–1.67) | Mild ID (n = 104) vs. no ID (n = 8229): te(Hg) est df = 2.224, chi-square = 11.176, p = 0.007. Severe ID (n = 258) vs. no ID (n = 8229): te(As, Pb) (interaction term) est df = 7.018, chi square = 12.996, p = 0.037. Any ID (n = 514) vs no ID (n = 8229): te(As,Pb) (interaction term) est df = 1.368, chi square = 6.073, p = 0.006 te(Hg) est df = 1.641, chi square = 9.964, p = 0.006 | High |
| Onicescu et al. (44) South Carolina, United States | Spatial importance parameter hierarchical logistic regression modeling. Beta weights and parameters | (1) sex, GA maternal age, race, parity, smoking, age of housing, population density, (2) sex, GA, maternal age, race, and parity | Association between ID and arsenic and mercury concentration in soil during pregnancy, controlling for infant sex, GA, maternal age and race. | Beta weights and parameters: Arsenic: β1 (95% CI): 0.013 (0.00074, 0.038) Mercury: β1 (95% CI): 0.11 (0.021, 0.39) | Med | |
| Genetic or biological | ||||||
| Guo et al. (70) Western China | Deviations from Hardy-Weinberg equilibrium, differences in allele and genotype distribution. Odds ratios | None. | Single marker analysis showed a positive association of ID with rs225012 and rs225010. Particularly with rs225012, TT genotype frequency was significantly higher in ID cases than in controls. | TT genotype frequency (x2 = 9.18, p = 0.00246). Combination of rs225012 and rs225010 (x2 = 15.04, df 2, global p = 0.000549) rs225012 OR 1.491 (1.033–2.152) rs225010 OR 1.43 (1.006–2.031) | High | |
| Hurtado et al. (67) Florida, United States | Logistic regression Odds ratios | BW, maternal education, sex, race-ethnicity, maternal age, child’s age at entry to WIC program | Mild or moderate ID associated with anemia. | The effect of hemoglobin was significant for risk of mild-moderate ID: aOR 1.28 (1.05, 1.60). Mild-moderate ID risk: low education (aOR 11.94), normal education (aOR 8.32), VLBW (aOR 4.58), LBW (aOR 2.50). | High | |
| Jelliffe-Pawlowski et al. (66) California, United States | Poisson regression. Prevalence ratios (PR) | Sex, race-ethnicity, plurality, GA, BW, maternal age, education, birthplace, and parity. | Children with chromosomal and other structural birth defects at increased risk of ID by age 7 years. Children with non-chromosomal defects were at substantially increased risk for all levels of ID. | Risk of ID, chromosomal and birth defects: PR 26.80 (22.70–31.70). Down syndrome PR: 211.70 (171.30–261.50) Sex chromosome defects: PR 57.40 (23.70–138.60). Non-chromosomal PR: 11.10 (8.70–14.20) | Risk of ID, Down syndrome aPR 178.40 (141.00–225.70) Non-chromosomal aPR 8.90 (6.80–11.50) | High |
| Shaw et al. (39) California, United States | Logistic regression OR | Maternal race-ethnicity, and gender | Case and control infants had similar percentages of TT and CT genotypes with some differences by ethnicity. | Hispanic children: TT genotype aOR 1.90 (0.70–5.00) CT genotype aOR 2.60 (1.10–5.80) Males with TT genotype aOR 5.40 [1.80–16.30] | Low | |
*Quality rating: high (90%+), med (75–89%), low (<75%), where percent is calculated on the number of “Y” ratings for each of the JBI rating items per study type rating system. OR, odds ratio; RRR, relative risk ratios; POBW, proportion of optimal birth weight; AGA, appropriate-for-gestational-age; GA, gestational age; SES, socioeconomic status; ASD, autism spectrum disorder; ID, intellectual disability; CP, cerebral palsy; PAR, population attributable risks; BW, birth weight; BMI, body mass index; SGA, small for gestational age; HR, hazard ratio; LBW, low birth weight; GEE, generalized estimating equations; GDM, gestational diabetes; APH, antepartum hemorrhage; VLBW, very low birth weight; VSGA, very small for gestational age; PTB, preterm birth; VPTB, very preterm birth; VPA, valproic acid; AED, antiepileptic drugs; PGD, pre-gestational diabetes; UTI, urinary tract infection; As, arsenic; Pb, lead; Hg, mercury; df, degrees of freedom; PR, prevalence ratios.