Light in the Rational Treatment of Autism? Effects of Metformin on Steroid Hormones in a Patient with Polycystic Ovarian Syndrome (PCOS)

Background: Metformin is an effective treatment option for type 2 diabetes mellitus, and it is, to this day, the most prescribed oral antiglycaemic drug. Besides its effects mainly on mitochondrial activity, an off-label use came up as a pharmaceutical for subjects with a diagnosis of polycystic ovarian syndrome (PCOS) along with altered steroid hormone homeostasis. Besides these effects, even an influence on mood and social behavior was described, leading to the aim of this case report to elucidate the effects before versus after treatment with metformin on steroid hormones and social behavior. Methods: A female patient with diagnosed PCOS was analyzed three times for steroid hormone levels. The first analysis was performed before treatment; the second, after a period of 71 days with metformin at 2 × 500 mg; and the third, after a total of 144 days with metformin at 2 × 500 mg. Spot urine probes were taken in the morning for a combined gas chromatography−mass spectrometry (GC-MS), and the steroid levels were adjusted for creatinine excretion. A questionnaire on social behavior (Autism Spectrum Questionnaire) was administered before treatment and after 71 days. Results: A decrease in all the steroid hormones measured was detected after 71 and 144 days of treatment with metformin, being more pronounced after 144 days of treatment and highly significant (p < 0.001). Furthermore, in the untreated state, the class of corticosterone metabolites showed increased values compared to the female reference values for TH-11-DH-corticosterone, TH-corticosterone, and 5a-TH-corticosterone. In the class of estrogen metabolites, increased values compared to the reference values were detected for 17b-estradiol; in the class of 11-deoxycortisol metabolites, an increase in TH-11-deoxycortisol was detected. For the class of cortisol metabolites, increased values compared to the reference values were detected for cortisone, TH-cortisone, a-cortolone, b-cortolone, 20b-dihydrocortisone, cortisol, TH-cortisol, 5a-TH-cortisol, a-cortol, 20b-dihydrocortisol, and 6b-OH-cortisol. No increases in androgen metabolites were detected. Interestingly, weight decreased from 93.4 kg to 91.3 kg after 71 days and fell to 82.7 kg after 144 days of treatment. The skeletal muscle mass was 30.1 kg at the first visit, decreasing to 29.9 kg and to 27.5 kg. No significant difference in the social behavior score from baseline to after 71 days of treatment was detected. Discussion: Metformin improved the steroid hormone profiles from levels above the upper reference values to the middle of the reference values after 71 days and to the lower ends of the reference values after 144 days of treatment. This implies not only that metformin has an effect on steroid hormone levels, but in addition that the efficacy of the pharmaceutical seems to depend on the time interval from intake. To summarize, in this patient, steroid hormones were affected but social behavior was not. If no effect of metformin on social behavior exists, this must be supported by further cases.