Methylation

James et al. (2004) looked at metabolites of methylation in the blood of 20 autistic children and compared them with a control group of 33 healthy children.  They found that the autistic children had lower S-adenosylmethionine (SAM) and higher S-adenosylhomocysteine (SAH) compared to the control group (indicating impaired methylation), with the ratio of SAM to SAH being 50% lower than in the control group.  They also found lower total glutathione (tGSH) and higher oxidized glutathione (GSSG) in the autistic children compared to control (indicating higher oxidative stress), with the ratio of tGSH to GSSG being 70% lower than in the control group.  They went on to give eight of the autistic children folinic acid and betaine for three months, adding in methylcobalamin for the fourth month.  They found that the metabolite levels in these children changed to be similar to the control group following this intervention.  The authors note that there are four times as many autistic males as females in the general population.  14 of the 20 autistic children in this study were boys, so the higher prevalence among males was reflected in the study population.  However, the gender ratio of the control group is not provided.  A study by El-Maarri et al. (2007) found higher methylation in males than females by about 2%, and as much as 16% in one specific region, so I think it’s possible that gender could be a source of confounding bias in the James et al. (2004) study.  If the control group was evenly split between male and female, then the average methylation in the control group may have been slightly lower than it would have been if the control group had four times as many males as females.  If they had used matching to ensure the same gender ratio in the control group as in the autistic group, their results could actually have been strengthened because the differences they measured between the two groups could have been even greater.  Also, future studies trying to replicate the findings could end up with slightly different effect sizes if they used a different gender ratio in the control group.  Another strategy that could have been used to address this potential confounding bias would be to analyze the data separately for each gender.  A RCT on the intervention that they used with the eight children would be a good design for a future study, and they could look at the results for males separately from females. 

Zou et al. (2020) looked at amino acids in the blood of 70 autistic and 70 control group children in China.  In this case they did use matching so the gender ratio of the control group was the same as that of the autistic children.  Their objective was to begin work on establishing an objective biomarker that could be used in diagnosis of autism; currently it is diagnosed on the basis of reported symptoms, which can vary considerably among cases.  A blood screening could help to get treatment started earlier which could make a big difference in these kids’ lives.  They put all the participants on the same diet and moderate physical activity for a week before the blood test.  They found significant differences between the groups in eight of the amino acids they measured. Homocysteine was associated with severity of symptoms and was the only amino acid found to have high diagnostic value when taken alone.  Homocysteine is associated with oxidative stress and impaired methylation, so this result was consistent with the James et al. (2004) study.  The authors acknowledge the possibility of selection bias, because all of the participants were the same ethnicity.  This means that the results may not be generalizable to children of other ethnicities.  While it would have been difficult for these researchers to choose their participants at random from multiple countries and ethnicities, clinicians can look for similar studies that looked at other ethnicities and compare the results.  The fact that they disclosed the possibility of selection bias in their report was very helpful in that the reader is alerted to take this into consideration and look for further studies if the population they are working with is different from the study population.  The authors suggest that future research could take place to see if amino acid supplements could help with autism symptoms; a RCT would be a good design to look at this intervention. 

References:

El-Maarri, O., Becker, T., Junen, J., Manzoor, S. S., Diaz-Lacava, A., Schwaab, R., Wienker, T., & Oldenburg, J. (2007). Gender specific differences in levels of DNA methylation at selected loci from human total blood: a tendency toward higher methylation levels in males. Human Genetics, 122(5), 505. 

James, S. J., Cutler, P., Melnyk, S., Jernigan, S., Janak, L., Gaylor, D. W., & Neubrander, J. A. (2004). Metabolic biomarkers of increased oxidative stress and impaired methylation capacity in children with autism. The American Journal of Clinical Nutrition, 80(6), 1611–1617. 

Zou, M., Li, D., Wang, L., Li, L., Xie, S., Liu, Y., Xia, W., Sun, C., & Wu, L. (2020). Identification of Amino Acid Dysregulation as a Potential Biomarker for Autism Spectrum Disorder in China. Neurotoxicity Research, 38(4), 992–1000.