Table 2 Summary of systematic reviews of environmental risk factors for autism spectrum disorders

Maternal immigration [53], 2015

10 population-based

The adjusted results showed higher rates of ASD if children had migrant mothers (2.69%) compared with children of non-migrant mothers (0.91%). Six out of ten studies found that giving birth postmaternal migration significantly raised risks of ASD, with the exception of children born to Hispanic migrant mothers who had lower rates of ASD (compared to all other migrant groups)

A meta-analysis by another group supports these findings

Preconceptional chemical exposure [49], 2014

3 case-control

All studies showed an increased rate of ASD in children whose parents were exposed to chemicals

The studies were limited by retrospective design and small sample sizes

Pesticide exposure during pregnancy [49], 2014

1 retrospective case-control; 3 prospective cohort

All showed an increased rate of ASD in children with gestational exposure to chemicals including two prospective studies of organophosphates

The studies provide enough evidence to justify larger studies of pesticide exposure in ASD

Pesticide exposure during childhood [49], 2014

1 retrospective cross-sectional; 1 prospective cohort; 1 computer model

One study showed an association between organophosphate exposure; one between DDE exposure and ASD, and one between phosphine exposure and ASD

Air pollution during prenatal period [49], 2014

6 retrospective

All studies showed an increased rate of ASD associated with air pollution (with particulate matter in three studies and with NO2 in two studies)

A meta-analysis of four prospective European cohorts showed no association between air pollution and ASD

Air pollution during childhood [49], 2014

3 retrospective case-control

All studies showed some association between air pollution and ASD (association with NO2 is replicated in two studies)

Exposure to other chemicals [49], 2014

2 prospective cohort; 4 retrospective case-control

No effect in prospective studies; retrospective studies found some effects for solvents and other toxicants

Retrospective studies used questionnaires

Toxic waste sites [49], 2014

2 case series, ecological

Association of Superfund site with ASD in both studies

Water pollutant [49], 2014

2 ecological

No association between water chlorination and ASD in either studies

Heavy metals [49], 2014

8 ecological

7 studies showed association between Hg and ASD, one showed no association

Ecological studies are limited by indirect measures of exposure

In-house flooring [49], 2014

1 cohort

Polyvinyl chloride vs. wood floor increases the risk of ASD

Mercury levels [49], 2014

29 case-control

12 studies with elevation in at least one tissue in patients with ASD

Studies measuring levels of heavy metals were limited by cross-sectional design and small sample size. Furthermore, it is unclear how much cross-sectional measurements of metals in the body reflect the causal effect of such exposures on ASD risk.

Lead levels [49], 2014

25 case-control

11 studies with elevation in at least one tissue in patients with ASD

Cadmium levels [49], 2014

14 case-control

5 studies with elevation in at least one tissue in patients with ASD

Aluminum levels [49], 2014

11 case-control

3 studies with elevation in at least one tissue in patients with ASD

Arsenic [49], 2014

8 case-control

5 studies with elevation in at least one tissue in patients with ASD

Nickel [49], 2014

3 case-control

2 studies with elevation in at least one tissue in patients with ASD

Uranium levels [49], 2014

2 case-control

1 studies with elevation in at least one tissue in patients with ASD

Tin levels [49], 2014

1 case-control

1 studies with elevation in at least one tissue in patients with ASD

Heavy metals and severity [49], 2014

7 correlational studies

All studies showed some correlation between ASD severity and heavy metal levels (particularly Hg and Pb)

These studies suggest a dose-response relationship between metal exposure and ASD.

Heavy metals and chelation [49], 2014

12 uncontrolled studies

All showed improvement in symptoms of ASD following chelation treatment

The studies are limited by lack of the control group

Urine porphyrin studies [49], 2014

4 case-series; 7 case-control

All studies showed some association with severity, prediction, or oxidative stress

The studies are limited by retrospective nature

Solvents, pesticides, and PCB [49], 2014

2 case series; 2 case-control

Case series showed levels above normal adult range for children with ASD; case-control studies showed no association between PCB and ASD

The studies were limited by either lack of proper control group or small sample size

Phthalate [51], 2016

2 cohort; 3 case-control

Three studies showed association between phthalate exposure and ASD; two studies showed no association

PBDE [49], 2014

1 case-control

PBDE resulted in higher activation of immune response in patients with ASD

Valproate [37], 2014

1 population-based prospective; 2 prospective case series; 1 retrospective

All studies found significant and strong association between prenatal valproate exposure and ASD

The association between valproate and ASD is likely to be causal given the strength of association and evidence for biological plausibility

Folic acid [44], 2016

11 studies

The findings were inconsistent; some studies provide support for the beneficial effect of folic acid on risk of ASD, whereas others show the opposite effect

Observational studies were limited by self-report, whereas RCTs were limited by use of multivitamin

Camel milk [115], 2016

2 small short-term RCTs

Results showed beneficial effects on ASD symptoms

It is unclear whether this benefit is clinical and whether it extends to longer period of treatment

Casein and gluten-free diet [116], 2014

16 studies

Findings are inconsistent

Studies with positive findings were mostly of low quality

Omega-3 [43], 2011

2 RCTs

Studies found no evidence for beneficial effects of omega-3 on ASD symptoms

Assisted reproductive technologies [36], 2013

2 cohort; 5 case-control

3 studies showed positive association between ART and ASD, 3 studies showed no association, and 1 study showed a negative association between ART and ASD. The authors concluded that at present, no evidence supports the association between ART and ASD

There are several sources of heterogeneity such as study design, exposure definition, data source, and confounders.