Hyperandrogenism? Increased 17, 20-Lyase Activity? A Metanalysis and Systematic Review of Altered Androgens in Boys and Girls with Autism

Abstract

Introduction: There is increasing evidence that steroid hormone levels and, especially, androgen levels are elevated in autism. An overactivity of 17, 20-lyase with a higher production of the testosterone precursors dehydroepiandrosterone (DHEA) and androstenedione/androstenediol seems especially present in autism. Methods: An encompassing literature analysis was performed, searching for altered androgens in children with autism and using preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines. Included were all studies published before 31 March 2021 found using the following electronic databases: PubMed, Google Scholar, Cochrane Library, Scopus, and TRIP. Eight studies with boys and three studies with girls where steroid hormone measurements were performed from either plasma, urine, or saliva were found and analyzed. Analyses were performed for DHEA(-S/-C), androstenedione/androstenediol, and testosterone. Effect sizes were calculated for each parameter between mean concentrations for children with autism versus healthy controls. Results: Higher levels of androgens in autism were detected, with the majority of calculated effect sizes being larger than one. Conclusions: We found higher levels of the main testosterone precursors DHEA, androstenedione, and androstenediol, likely causing an additionally higher level of testosterone, and an increased 17, 20-lyase activity is therefore implied. Medications already used in PCOS such as metformin might be considered to treat hyperandrogenism in autism following further research.

Keywords: androgens; extreme male brain theory of autism; steroid hormones.

Figure 1

Steroid hormone activity—higher 17, 20-lyase activity should be indicated by higher levels of DHEA, DHEA-S and androstenedione. One step further, higher levels of androstenediol and testosterone in line with increased 17B-HSD and 3B-HSD are indicative of a general hyperandrogenism in autism [39].

Figure 2

Flow diagram of meta-analysis procedure.

Figure 3

(a–c) Random effects models with significance level, Cochrane Q, I², and effect sizes of analyzed metabolites in boys with autism versus healthy controls. (a) Androstenedione/androstenediol indicated with *; (b) DHEA/DHEA-S indicated with *, DHEA-C indicated with **; (c) testosterone/free testosterone indicated with *.

Figure 3

(a–c) Random effects models with significance level, Cochrane Q, I², and effect sizes of analyzed metabolites in boys with autism versus healthy controls. (a) Androstenedione/androstenediol indicated with *; (b) DHEA/DHEA-S indicated with *, DHEA-C indicated with **; (c) testosterone/free testosterone indicated with *.

Figure 4

(a–c) Random effects models with significance level, Cochrane Q, I², and effect sizes of analyzed metabolites in girls with autism versus healthy controls. (a) Androstenedione/androstenediol indicated with *; (b) DHEA; (c) testosterone/free testosterone indicated with *.

Figure 4

(a–c) Random effects models with significance level, Cochrane Q, I², and effect sizes of analyzed metabolites in girls with autism versus healthy controls. (a) Androstenedione/androstenediol indicated with *; (b) DHEA; (c) testosterone/free testosterone indicated with *.

Figure 5

Funnel plots for boys: x-axis—standardized mean difference; y-axis—standard error for (a) androstenedione/androstenediol, (b) DHEA-(C/S), and (c) (free) testosterone.

Figure 6

Funnel plots for girls: x-axis—standardized mean difference; y-axis—standard error for (a) androstenedione/androstenediol, (b) DHEA-(C/S), and (c) (free) testosterone.