Studies demonstrate that FMT significantly improves behavioral symptoms of ASD. Two open-label clinical trials and two retrospective studies found that post-intervention with FMT, ASD severity was significantly reduced relative to baseline when measured using the Childhood Autism Rating Scale (CARS) (Kang et al., 2017; Kang et al., 2019; Li et al., 2021; Pan et al., 2022; Zhang et al., 2022). The CARS is conducted by a professional evaluator and has previously been reported as a valid diagnostic tool with good sensitivity and specificity (Chlebowski et al., 2010). Kang et al. found that at both the 8-week and 2-year follow-ups there were no signs of reversal of the improvements to ASD symptoms, (Kang et al., 2017; Kang et al., 2019); 2 years after intensive treatment with FMT, only 17% children were rated as having severe ASD, compared with 83% at baseline, and 44% children scored below the ASD diagnostic threshold (Kang et al., 2019). This demonstrates the potential of FMT as a therapeutic with long-term benefits for children with severe ASD.

In contrast, Li et al. found that improvements in social skills deficits post-intervention were reversed at the 8- to 12-week follow-up in an open-label study of 40 children with ASD and GI problems (Li et al., 2021). A possible explanation for the differences in the findings is that different doses and treatment regimens were used. Kang et al. opted for a high initial dose of FMT followed by daily maintenance doses for 8 weeks, whereas Li et al. delivered a high dose once a week for 4 weeks (Kang et al., 2017; Li et al., 2021). Figure 1 shows the treatment regimens used in human trials. Two additional studies also demonstrated that the effect of FMT on the severity of ASD symptoms increased with increasing treatment courses (Pan et al., 2022; Zhang et al., 2022). This evidence suggests that prolonged treatment, or multiple courses, may achieve longer-lasting benefits for children with ASD.

Overall, there were no reported severe adverse reactions to FMT, which suggests that it is a safe and well-tolerated treatment. However, Kang et al. found that 5%, 39%, and 28% of participants suffered from mild to moderate rashes, hyperactivity, and tantrums/aggression, respectively, when treated with vancomycin treatment prior to FMT (Kang et al., 2017). The use of preparation procedures in the studies varied (see Figure 1 ). The preparation procedures for FMT often include vancomycin, proton pump inhibitors, and bowel cleansing to prepare the intestinal landscape for engraftment of the donor microbiota (Kang et al., 2017). The effects of preparatory procedures on the safety and efficacy of FMT must be considered as previous research highlights the role of vancomycin and proton pump inhibitors on recipient microbiota, but they can result in negative side effects for participants (Vrieze et al., 2014; Freedberg et al., 2015; Sandler et al., 2016). Future research should compare the effects of FMT with and without preparatory procedures to determine whether it is necessary for FMT success. No differences on treatment safety or therapeutic outcome have been noted between receiving doses of FMT via oral or rectal administration in studies that used both administration routes (Kang et al., 2017; Li et al., 2021). Therefore, the ability to swallow capsules and patient/parent preference should be the determining factors when deciding between administration routes.

Two rodent studies found that social deficits in ASD mice were normalized upon treatment with FMT (Goo et al., 2020; Li et al., 2020); another three studies using rodents found that there was an increasing trend in social behavior upon treatment with FMT, although these findings did not reach significance (Chen et al., 2020; Abuaish et al., 2021; Abuaish et al., 2022). While these findings support the results from human trials, a criticism of using chemical- or immune-activated rodent models of ASD is that they are unlikely to simulate the complex interplay between pathophysiology, genetics, and environmental factors that humans experience. This is supported by Bronfenbrenner’s ecological systems theory, which describes the effect of the external environment on a child’s development, including in those children with developmental disorders (Bronfenbrenner, 1979).

FMT significantly improves GI symptoms for children with ASD (Kang et al., 2017; Kang et al., 2019; Li et al., 2021; Pan et al., 2022; Zhang et al., 2022). Indigestion, diarrhea, abdominal pain, and constipation were all significantly improved relative to baseline across multiple studies, alongside a reduction in the number of days where participants experienced abnormal or no stools (Kang et al., 2017; Kang et al., 2019; Li et al., 2021; Pan et al., 2022). These improvements persisted at the 2-year follow-up (Kang et al., 2019). The presented findings are not surprising considering the evidence for the use of FMT to treat isolated GI health conditions (Borody, 2003; Kunde et al., 2013; Allegretti et al., 2021).

The Gastrointestinal Symptom Rating Scale (GSRS) is a disease-specific questionnaire based on 15 elements, divided into five symptom groups that describe GI symptoms such as reflux, abdominal pain, indigestion, diarrhea, and constipation. According to the GSRS, symptoms are rated from one to seven on a Likert-type scale, where one represents no troublesome symptoms and seven represents very distressing symptoms (Kulich et al., 2008). Kang et al. found a significant positive correlation between the percentage changes in GSRS scores and CARS scores (Kang et al., 2019). This is supported by previous research that shows a link between GI and ASD severity, but from these data it was not possible to ascertain a causative effect (Babinska et al., 2020). In addition, Zhang et al. found a greater positive effect on ASD severity in children who were not constipated than for those who were constipated prior to FMT (Zhang et al., 2022). The authors suggested that, in children who have pre-existing GI symptoms, FMT may better improve ASD symptom severity if the children’s GI problems are treated first. Apart from one study, all human trials used only participants who had co-occurring GI issues alongside ASD (see Table 1 ). This creates a gap in the research, with more evidence required to evaluate whether ASD symptoms would improve to a greater extent if there were no co-occurring GI symptoms.

Both human and rodent studies show that treatment with FMT leads to an increase in microbial diversity and richness relative to the baseline in participants with ASD (Kang et al., 2017; Kang et al., 2019; Chen et al., 2020; Abujamel et al., 2022). This is of importance because research suggests that higher levels of gut microbiota diversity is considered healthier due to stronger microbiota integrity and functional redundancy, which helps to protect the gut from pathogenic bacteria (Larsen and Claassen, 2018). Furthermore, FMT was able to promote engraftment of donor microbiota so that participants’ microbiota composition was nearly indistinguishable from that of the TD donors 8 weeks post-intervention (Kang et al., 2017; Li et al., 2021). Interestingly, at the 2-year follow-up, the participants’ microbiota was again distinguishable from the donors, but the features of a healthy gut, such as a high level of bacterial density and a high abundance of Prevotella, were retained, as illustrated in Figure 2 (Kang et al., 2019). This suggests that FMT leads to a long-term benefit in gut health by shifting the features of gut microbiota toward that of TD controls.

Specific differences in gut microbiota composition and abundance of species were unique to each study. For example, Kang et al. found that, following FMT, the relative abundances of Bifidobacterium, Prevotella, and Desulfovibrio were all significantly higher than baseline at both 8 weeks and 2 years post-intervention (Kang et al., 2017; Kang et al., 2019), whereas Chen et al. found that there was a “decrease in the relative abundance of Prevotella_other and a significant increase in the relative abundance of Oscillospira” (Chen et al., 2020). In rodent models, contrasting findings may be due to the different models used to induce ASD behaviors; using different chemical or immune-activated models will inherently change the structure of the gut microbiota, as shown by previous research (Coretti et al., 2017). In human studies, the variation in findings may be attributed to differences between the intestinal landscapes of both donors and FMT recipients. Li et al. found bacterial differences prior to treatment in participants who responded to FMT compared with those who saw no improvement post-intervention (Li et al., 2021). There is also evidence that indicates that the success of FMT depends on the microbial composition of donor samples, and there may be so-called “super-donors” who have the optimum microbial composition for a therapeutic intervention (Kump et al., 2018). This evidence combined suggests that there is a future where donors and recipients may be matched based on their gut microbiota to enable restoration of specific microbial deficits. Collection of useful clinical information, such as the initial microbiota composition, will help with the characterization of recipients and donors to reduce the variability of patient responses in the future and optimize targeting of donors.

To date, there has been no blinded randomized control trial (RCT) to investigate the effects of FMT on ASD symptoms. In order to reduce the possibility of the placebo effect and researcher bias, a placebo-controlled, double-blind RCT would be of value to the evidence base. The sample should be carefully considered as there is a well-documented sex bias in ASD diagnosis and research (Tillmann et al., 2018). All included studies report vastly higher proportions of male participants than female (see Table 1 ). Evidence shows that males and females present symptoms of ASD differently, so it is crucial that research into therapeutic options is carried out to reflect this (Harrop et al., 2018; Tillmann et al., 2018). Furthermore, the studies all used relatively small sample sizes of fewer than 60 participants: a larger sample size will increase the statistical power of any findings from a future RCT. Finally, the included studies only test up to the age of 18 years. As previously discussed, symptoms of ASD often persist into adulthood (Chamak and Bonniau, 2016). Therefore, it is important to consider whether the symptoms can be ameliorated in older individuals to reduce the significant negative outcomes they experience. There are several strengths of this review, including that it summarizes research studies that have shown FMT to be effective in alleviating symptoms of ASDs, while at the same time discussing limitations and future directions.