Onto this paper of Shizuku's: www.frontiersin.org/articles/10.3389/fpsyg.2016.01053/full
I still can't decide if Shizuku actually reads and understands this stuff or is just spitting them out using a keyword search. Shizuku is quite good at finding things which say "no effect", the problem is they're almost always not what we're talking about.
Maybe this works with people who don't click on links or say things like "Trust The Science"?
Anyway, that paper was studying girls who underwent treatment for precocious puberty. NOT girls having their normal development blocked for "gender correction".
What we're primarily worried about here is interfering with the normal development that happens leading up to adulthood. Any such problem would be a lot harder to spot in a precious puberty treatment. I'm not even sure we'd know what we're looking for at that point. One issue the authors note:
Since a [randomized control trial] cannot be conducted due to ethical reasons ... [but we can give them to people for gender correction without having done any such trials!!!!] ... ...comparison of treated CPP children and controls matched for either chronological or biological age. [...] Furthermore, the question remains if matching should be done by chronological or biological age. 
They went with chronological, which is the most likely to find no effect, but what if what you've just done is blocked a year of their normal development because it's "early"? What if these precocious puberty children should be gaining brain function now - this is their time - and won't catch up everything later? Could be a problem if they still end puberty early despite the blockers.
Anyway, we know there's been quite a lot of studies on younger children having shorter-term regimes. What we're looking for is cognitive studies of children having their puberty totally blocked all the way up to a cross-sex hormone regime (which that bone density study did manage to do). That's a totally different ball-game to a short period of blocking precocious puberty at 9.
Doing my own cherry-picking now (if Shizuku can do it on this thread, only fair to do it back) - there was one potential negative finding in that precocious puberty study:
Only in the Trail Making Test—Number Sequencing, assessing processing speed, the CPP group showed significantly poorer performance (Table 2). This finding is difficult to explain since neither the very similar Trail Making Test—Letter Sequencing, nor any other of the processing speed tests showed significant differences between the groups. Taking into account that the p-values were not corrected for multiple testing, it is possible that this finding is accidental. In line with this, the CPP girls' parents did not report any problems with regard to executive functioning as measured by the BRIEF questionnaire.