Randomised controlled trial of simvastatin treatment for autism in young children with neurofibromatosis type 1 (SANTA).

Only 30 children were randomised and 26 completed endpoint, so the study was very small. The trial was not powered for formal behavioural treatment effects. Responder counts were tiny and not statistically tested. Two participants in each arm were on stimulant medication, so behaviour cannot be attributed only to simvastatin in those children. Dosing was based on statin use in other conditions, not on dose-finding for NF1-autism. The trial lasted 12 weeks, so longer-term effects are unknown. Peripheral MAPK is not a confirmed marker of brain Ras activity in human NF1. Many imaging analyses had missing or incomplete data because of movement artefacts and scanning challenges in young children. Several imaging findings were uncorrected for multiple comparisons or changed after baseline adjustment. Baseline differences existed, including lower deep grey nuclei Glx in the treatment group and different proportions of inherited NF1 mutations. The default mode network comparison was null at the 5% corrected level, while machine-learning findings were exploratory.

Funded by Central Manchester University Hospitals NHS Foundation Trust. The supplied text does not state any additional conflicts of interest.

Do not read the 3 of 12 statin responders versus 0 of 14 placebo responders as evidence that simvastatin improves autism symptoms. The trial was not powered for behavioural effects, no significant between-group behavioural differences were seen, and the authors state that the study does not support clinical use of simvastatin in these children at this time. Two participants in each arm were also on stimulant medication, so behaviour cannot be attributed only to simvastatin in those children.