"These results provide the first evidence that carbamazepine in drinking water and at typical environmental concentrations is transmitted from mother to embryo."
So said the findings reported by Gaurav Kaushik and colleagues  who continued a research theme looking at the possibility of an environmental 'exposome' also potentially overlapping with [some] autism (see here). The autism connection came from the a paper by Thomas & Klaper  who previously reported that putting fish in a tank of water containing various unmetabolized psychoactive pharmaceuticals (UPPs) including carbamazepine resulted in some changes to gene expression patterns "associated with idiopathic autism."
This time around the idea was to test whether "psychoactive pharmaceuticals (fluoxetine, venlafaxine, and carbamazepine) administered through the drinking water at environmental concentrations to pregnant mice could reach the brain of the developing embryo by crossing intestinal and placental barriers." Authors added deuterated standards (2H-isotope labelled) of the pharmaceuticals in question to the water so they were able to quantify rat embryo brain and liver levels of them in comparison with control animals. As it turns out, carbamazepine (normally used to treat epilepsy) was the only pharmaceutical detected - "using isotope ratio mass spectrometry" - in those developing embryos and led authors to that opening sentence. Combined with the previous work they suggest that "carbamazepine may be associated with ASD [autism spectrum disorders] in infants" and as such "warrant the closer examination of psychoactive pharmaceuticals in drinking water and their potential association with neurodevelopmental disorders."
There is an obvious 'shock' value to this research insofar as the suggestion that minute concentrations of certain pharmaceuticals fairly commonly found in drinking water might have repercussions when it comes to the 'risk' of offspring autism. On that basis I tread carefully in the discussion of this topic.
I should initially tackle the idea that as much as we would like to think our drinking water comes from some untouched crystal clear spring on top of a mountain, the reality is perhaps a little different. Drinking water is mostly of a pretty good standard (well, mostly) but that's not to say that some of our population habits don't have an effect on its quality. With carbamazepine specifically in mind, there is quite a body of research suggesting that very low levels of this and other medicines do turn up in drinking water  across various different geographies and are pretty stable in said waters  at least during certain points of water purification.
The issue however is one of translating the Kaushik results from lab to real-life. Not only that they report results based on fish and mouse models but also that a wide variety of prescription and perhaps even non-prescription compounds may turn up in real-life drinking water and could have many different (often opposing) effects when it comes to things like gene expression linked to various diagnoses. Combined with the idea that gene expression patterns 'linked to idiopathic autism' does not necessarily mean that it is linked to all idiopathic or other kinds of autism and as you might see, we have some way to go yet before anyone talks about water supplies increasing risk of autism or anything else. As far as I am aware, behaviours pertinent to autism were also not analysed as part of the current research in this area bearing in mind animal models are not necessarily a true reflection of any condition.
Having said all that, I do think there is more to do in this area - first and foremost with a view to independent replication of these results. I'd also like to see a little more information about the mechanisms that might be involved aiding the passage of any UPPs to the embryo and onwards the genetic (and epigenetic) effects that they may be linked to, again taking into account other work from some of the researchers discussed today . We then get to possibility of some modelling with other animal sets and then studies towards how such effects could be examined with people in mind (not forgetting that water is used for quite a few purposes other than just drinking).
 Kaushik G. et al. Maternal exposure to carbamazepine at environmental concentrations can cross intestinal and placental barriers. Biochem Biophys Res Commun. 2016 Apr 19. pii: S0006-291X(16)30594-0.
 Thomas MA. & Klaper RD. Psychoactive pharmaceuticals induce fish gene expression profiles associated with human idiopathic autism. PLoS One. 2012;7(6):e32917.
 Wu M. et al. Occurrence and fate of psychiatric pharmaceuticals in the urban water system of Shanghai, China. Chemosphere. 2015 Nov;138:486-93.
 Cormier G. et al. The degradation behaviour of nine diverse contaminants in urban surface water and wastewater prior to water treatment. Environ Sci Process Impacts. 2015 Dec;17(12):2051-65.
 Kaushik G. et al. Psychoactive pharmaceuticals as environmental contaminants may disrupt highly inter-connected nodes in an Autism-associated protein-protein interaction network. BMC Bioinformatics. 2015;16 Suppl 7:S3.
Kaushik G, Huber DP, Aho K, Finney B, Bearden S, Zarbalis KS, & Thomas MA (2016). Maternal exposure to carbamazepine at environmental concentrations can cross intestinal and placental barriers. Biochemical and biophysical research communications PMID: 27105911