Scrutable

A note in The Conversation UK relates that

It is possible to improve symptoms in autistic children with a cheap generic drug... The drug, bumetanide, is widely used to treat high blood pressure and swelling...

We were motivated to test bumetanide as a result of background findings which suggested that the drug changed important brain chemicals in mouse models of autism; and also by some studies, including in autistic teenagers, showing that bumetanide may have beneficial effects.

There were 81 children with moderate to severe autism in our study – 42 in the bumetanide group, who received 0.5mg of bumetanide twice a day for three months; and 39 children in the control group, who received no treatment. The children were three to six years of age.

Some of the children had their brains scanned using magnetic resonance spectroscopy (MRS) – 38 in the bumetanide group and 17 in the control group. MRS is a non-invasive way of measuring chemicals in the brain. For our study, we measured brain chemicals called GABA and glutamate, which are important for learning and brain plasticity (the brain’s ability to change and adapt as a result of experience).

In the bumetanide group, autism symptoms improved as measured by the childhood autism rating scale (CARS) and also by a doctor’s overall impression. The doctors who were assessing symptom change were “blind” to treatment – that is, they were unaware of who was receiving bumetanide. Improvements in symptoms were associated with changes in the brain chemicals GABA/glutamate ratios and, in particular, with decreases in GABA.

Very small study etc. but may point towards something useful?

Zhang et al. Symptom improvement in children with autism spectrum disorder following bumetanide administration is associated with decreased GABA/glutamate ratios Translational Psychiatry volume 10, Article number: 9 (2020)

Open Access: https://www.nature.com/articles/s41398-020-0692-2

Hmmm. If I do a basic ANOVA and compute the interaction term for "Total score", using the values from the last-but-1 line of Table 1 and the first line of Table 2 (means of (38.15, 37.40, 37.27, 34.51) and corresponding SDs), I get F(1,81)=3.08, p=.08. For "Num.item >= 3" using the same method I get F(1,81)=2.52, p=.11. But the authors used a complex mixed-effects model, which let to far more impressive p-values by the time they had thrown in a selection of covariates. I wonder if that analysis plan was preregistered, and if not, how many goes it took them to get their model right?

Interesting one there - thank you. Will have a fuller read later.

However, one thing which jumped straight out at me (Me? With my background?) was the use of CARS as their assessment tool. There are a couple of problems with CARS, especially its false positives from identifying non-autistic, but learning disabled, bairns as autistic, and difficulties in separating ASD from the broader PDD -NOS groupings - see here for example.

I've encountered very few papers on autism which used what we used (the NICE "gold standard") - ADOS, ADI-R, multiple observations in different settings by different professionals, full, separate from rating scales, developmental history, possible use of WISC-R to eliminate other cognitive issues. I can appreciate the utility of using shorter, quicker assessment tools as opposed to the full diagnostic battery a CAMHS team would use, but the limitations and errors this causes need to be fully acknowledged.

Honest guv, I did do much more (and preferred to and was more interested in) loadsa other things than ASD assessment, but I seem to post more about ASD than any other area of my old work...

Oh and anyone citing mouse models can piss right off!

Couple of other points (after a quick read of both links) - at least all the bairns had a decent ASD diagnosis (ADI-R, ADOS etc, decent team involved, DSM-5 criteria), which is not always the case in papers on "ASD".

Steamy's point about the stats is a good one - it all has a look of a pistol being held to the data's head and it being asked if there is anything else it wants to say...

I'll be interested to see if this is replicated by other teams, especially ones which don't rely on CARS.

murmur wrote: ↑

Wed Jan 29, 2020 11:53 am

Steamy's point about the stats is a good one - it all has a look of a pistol being held to the data's head and it being asked if there is anything else it wants to say...

Actually it isn't. I made a very basic mistake of performing a 2x2 between-subjects ANOVA when in fact the design is 2x2 mixed (2 groups, pre-post), so it's not clear that any torturing is going on. I do wish they had published the results of the simple 2-between, 2-within ANOVA though.

The effect sizes are huge (eta squared of 0.61 and 0.26 by my calculation), but of course it's possible that the medicine is actually having a strong effect on most participants.

I remain skeptical in principle that blood pressure tablets can help autism symptoms, also given the hype about anything to do with autism, but I withdraw my initial statistical skepticism and will await other opinions and/or future research.

My gut instincts are that a) there are diagnostic difficulties with working patients that are aged 3-6 because of the impact of differing rates of brain development, and my instinct that a fair amount of low level social behaviour develops around this time in childhood, and b) that there's potentially an opportunity for drugs to modulate that phase of brain development, and achieve big effects. This means that a) I'm initially sceptical, and b) wouldn't be as astonished if this turns out to be correct as I would be if they were claiming the same results in adults.

Of course, my guts are mostly full of sh.t and aren't well read around either child development or autism. But would anyone like to use their brain to comment on those thoughts?

As a father of an autistic child and a long time teacher of autistic children, I'm suspicious if only because of all the sh.t science that autism gets thrown its way. In fact autism could be used as an archetypal case for studies on sh.t science in medicine and education.

Then of course this new research will be criticised by the autism-is-not-a-disease-that-needs-to-be-treated-they-are-different-not-sick brigade.

sTeamTraen wrote: ↑

Wed Jan 29, 2020 2:47 pm

murmur wrote: ↑

Wed Jan 29, 2020 11:53 am

Steamy's point about the stats is a good one - it all has a look of a pistol being held to the data's head and it being asked if there is anything else it wants to say...

Actually it isn't. I made a very basic mistake of performing a 2x2 between-subjects ANOVA when in fact the design is 2x2 mixed (2 groups, pre-post), so it's not clear that any torturing is going on. I do wish they had published the results of the simple 2-between, 2-within ANOVA though.

The effect sizes are huge (eta squared of 0.61 and 0.26 by my calculation), but of course it's possible that the medicine is actually having a strong effect on most participants.

I remain skeptical in principle that blood pressure tablets can help autism symptoms, also given the hype about anything to do with autism, but I withdraw my initial statistical skepticism and will await other opinions and/or future research.

It seems that bumetanide for autism in children is apparently not a new idea; here's one of their references from 2012 which cites a pilot study from 2010. The tl;dr is that (and this is very reductionist) it's down to chloride ions and γ-aminobutyric acid (GABA):

Brain maturation is associated with a developmental sequential expression of voltage gated, receptor synapse–driven channels and brain patterns (5, 6). The developmental shifts of the actions of the inhibitory transmitter GABA is but one example of these changes. Immature neurons have a higher (Cl−)I than adults leading to paradoxical excitatory actions of GABA (6). This is because of an early expression of the co‐transporter NKCC1 that imports chloride and a late operation of KCC2 that export chloride form neurons (7). In addition, the regulation of (Cl−)I is altered by a variety of insults, lesions, seizures and neurological disorders thereby converting the actions of GABA from inhibitory to excitatory (8, 9). Consequently, diuretic agents that reduce (Cl−)I constitute novel antiepileptic and neuro‐protective agents (10, 11) and are currently being tested in large clinical trials in infantile epilepsies (Nemo FP7 EU program and Harvard‐based trial).

The references are

5 Ben‐Ari Y, Gaiarsa JL, Tyzio R, Khazipov R. GABA: a pioneer transmitter that excites immature neurons and generates primitive oscillations. Physiol Rev 2007; 87: 1215– 84.
6 Spitzer NC, Gu X, Olson E. Action potentials, calcium transients and the control of differentiation of excitable cells. Curr Opin Neurobiol 1994; 4: 70– 7.
7 Rivera C, Voipio J, Payne JA, Ruusuvuori E, Lahtinen H, Lamsa K, et al. The K+/Cl− co‐transporter KCC2 renders GABA hyperpolarizing during neuronal maturation. Nature 1999; 397: 251– 5.
8 Cohen I, Navarro V, Clemenceau S, Baulac M, Miles R. On the origin of interictal activity in human temporal lobe epilepsy in vitro. Science 2002; 298: 1418– 21.
9 Khalilov I, Le Van QM, Gozlan H, Ben Ari Y. Epileptogenic actions of GABA and fast oscillations in the developing hippocampus. Neuron 2005; 48: 787– 96.
10 Dzhala VI, Talos DM, Sdrulla DA, Brumback AC, Mathews GC, Benke TA, et al. NKCC1 transporter facilitates seizures in the developing brain. Nat Med 2005; 11: 1205– 13.
11 Nardou R, Ben‐Ari Y, Khalilov I. Bumetanide, an NKCC1 antagonist, does not prevent formation of epileptogenic focus but blocks epileptic focus seizures in immature rat hippocampus. J Neurophysiol 2009; 101: 2878– 88.

Boustrophedon wrote: ↑

Wed Jan 29, 2020 3:30 pm

As a father of an autistic child and a long time teacher of autistic children, I'm suspicious if only because of all the sh.t science that autism gets thrown its way.

This seems like a big problem, kind of how if Jesus showed up tomorrow he wouldn't get a visa for most of the countries where he'd be needed.

sTeamTraen wrote: ↑

Wed Jan 29, 2020 4:37 pm

Boustrophedon wrote: ↑

Wed Jan 29, 2020 3:30 pm

As a father of an autistic child and a long time teacher of autistic children, I'm suspicious if only because of all the sh.t science that autism gets thrown its way.

This seems like a big problem, kind of how if Jesus showed up tomorrow he wouldn't get a visa for most of the countries where he'd be needed.

And definitely not the US.

sTeamTraen wrote: ↑

Wed Jan 29, 2020 2:47 pm

murmur wrote: ↑

Wed Jan 29, 2020 11:53 am

Steamy's point about the stats is a good one - it all has a look of a pistol being held to the data's head and it being asked if there is anything else it wants to say...

Actually it isn't. I made a very basic mistake of performing a 2x2 between-subjects ANOVA when in fact the design is 2x2 mixed (2 groups, pre-post), so it's not clear that any torturing is going on. I do wish they had published the results of the simple 2-between, 2-within ANOVA though.

The effect sizes are huge (eta squared of 0.61 and 0.26 by my calculation), but of course it's possible that the medicine is actually having a strong effect on most participants.

I remain skeptical in principle that blood pressure tablets can help autism symptoms, also given the hype about anything to do with autism, but I withdraw my initial statistical skepticism and will await other opinions and/or future research.

Aaaaah bugger!

Need to go away and brush up on my own stats knowledge, which is verey rusty...

Re your second point: given that the whole idea of a diuretic having that sort of neurological effect came from rodent "models" I would regard your scepticism of that point as justified. I remain highly unconvinced of the validity of rodent "models" of "autism" and will stay that way until someone comes up with something VERY convincing that these are in anyway analagous to a highly complex condition of a far more complex brain.

dyqik wrote: ↑

Wed Jan 29, 2020 2:55 pm

My gut instincts are that a) there are diagnostic difficulties with working patients that are aged 3-6 because of the impact of differing rates of brain development, and my instinct that a fair amount of low level social behaviour develops around this time in childhood, and b) that there's potentially an opportunity for drugs to modulate that phase of brain development, and achieve big effects. This means that a) I'm initially sceptical, and b) wouldn't be as astonished if this turns out to be correct as I would be if they were claiming the same results in adults.

Of course, my guts are mostly full of sh.t and aren't well read around either child development or autism. But would anyone like to use their brain to comment on those thoughts?

What I called "barn door obvious" autism can be spotted age 3-6, more Aspergery bairns not really, some others, especially female, definitely not (clinical practice experience).

This is the nature of the beast: when developmental delays and problems are involved you do need to wait to see if development has actually taken place or not...

I'd like to see some replication in an older population, whether teenage or adult doesn't bother me.

Apparently bumetanide is also being investigated as a possible anti-epileptic 'cos "n the brain, bumetanide blocks the NKCC1 cation-chloride co-transporter, and thus decreases internal chloride concentration in neurons. In turn, this concentration change makes the action of GABA more hyperpolarizing, which may be useful for treatment of neonatal seizures, which quite often are not responsive to traditional GABA-targeted treatment, such as barbiturates." From Wikipedia.

murmur wrote: ↑

Wed Jan 29, 2020 5:44 pm

dyqik wrote: ↑

Wed Jan 29, 2020 2:55 pm

My gut instincts are that a) there are diagnostic difficulties with working patients that are aged 3-6 because of the impact of differing rates of brain development, and my instinct that a fair amount of low level social behaviour develops around this time in childhood, and b) that there's potentially an opportunity for drugs to modulate that phase of brain development, and achieve big effects. This means that a) I'm initially sceptical, and b) wouldn't be as astonished if this turns out to be correct as I would be if they were claiming the same results in adults.

Of course, my guts are mostly full of sh.t and aren't well read around either child development or autism. But would anyone like to use their brain to comment on those thoughts?

What I called "barn door obvious" autism can be spotted age 3-6, more Aspergery bairns not really, some others, especially female, definitely not (clinical practice experience).

This is the nature of the beast: when developmental delays and problems are involved you do need to wait to see if development has actually taken place or not...

I'd like to see some replication in an older population, whether teenage or adult doesn't bother me.

Well there's only "A more recent report revealed that bumetanide could reduce amygdala activation in adolescents with ASD, in which authors hypothesized that bumetanide restores the E-I balance of brain 19." in the OP article so far.

19. Hadjikhani, N. et al. Bumetanide for autism: more eye contact, less amygdala activation. Sci. Rep. 8, 3602 (2018). https://doi.org/10.1038/s41598-018-21958-x

I wrote to the corresponding authors of this article on 31 January asking for a suitably-anonymised copy of the dataset, signing off with my doctorate and affiliation. I have received no reply so far.

I'm prepared to bet the person opening the letter/email said "Oh f.ck, he's taken an interest."

JQH wrote: ↑

Thu Feb 27, 2020 8:44 pm

I'm prepared to bet the person opening the letter/email said "Oh f.ck, he's taken an interest."

Probably true.