Skip to Content

Ultrasound For Brain Drug Delivery – Not So Fast

I wrote a couple of years ago here about the idea of making the blood-brain barrier more permeable by the use of focused ultrasound, in the presence of injected microbubbles. This would be a very useful thing if it works – as anyone who’s been concerned with central nervous system drugs (or drug delivery in general) well appreciates, the lining of the cerebral blood vessels is tight enough to send a lot of potentially useful therapies bouncing off it. The brain is a well-protected organ, and all sorts of schemes have been hatched over the years to get compounds past that last line of defense.

Since writing that post, I’d been wondering how the ultrasound idea had been working out, so I was particularly interested to see a recent exchange in the pages of PNAS. This paper (from a team of NIH researchers) is a detailed study of rat brain tissue after exposure to the technique, and they find a variety of markers of injury and inflammation at the site of ultrasound exposure. In fact, they go as far as to say that the entire process is hard to distinguish from ischemia or head trauma, and advocate caution before any further use of it in human clinical trials.

That’s a bucket of cold water, for sure, and the paper drew a published response from a team of researchers in Toronto who have been working in this area for years now. They suggest that the NIH paper is wrongly interpreting their results – that there are several markers of ischemic damage, etc., that are not induced by the ultrasound treatment, that no cell death is noted (and no scar tissue, etc.), and that the inflammatory markers resolve within hours to days. Buried inside their response is a note that the NIH group, they say, used five to ten times the recommended concentration of microbubbles.

That one drew a further response from the NIH team. The entire exchange, I should note, is conducted in flawless high academic style – the Toronto group thanks the NIH people for publishing their results, which they say confirm their own work (before then getting down to the disagreements, in the unsaid “If you’d only done this properly” style). The NIH group in turn thanks the Toronto one for their response, which they say now gives them a chance to expand on the implications of their own work (in the unsaid “If you’d only understood what we wrote” style). They say that the Toronto reply misses out on a number of details of the microbubble preparation and the ultrasound technique used, and that they did not, in fact, use any sort of large excess of microbubbles, and made sure that they were well below the ultrasound levels that are known to cause tissue damage on their own, etc. They also point out that if this technique is actually going to be used in human trials for neurodegenerative diseases, that the patients involved will have to have much larger areas of their vasculature sonicated (just to get enough drug through).

So there the matter stands. At the very least, everyone can agree that (as the phrase goes) “more study is needed”. Who will be vindicated at the end of that study is too early to say, but given the (appropriate) caution needed when messing with the blood vessels around someone’s brain, I’d say that it might be a while before we see this technique applied. . .

35 comments on “Ultrasound For Brain Drug Delivery – Not So Fast”

  1. Adam says:

    Thanks for highlighting this. Just one comment on your last sentence, this technique is already being applied. See:

    1. Derek Lowe says:

      Are these recruiting, or have they already moved on to treating patients?

      1. Adam says:

        I don’t know for sure but they’ve been listed as recruiting for quite some time so I imagine they have treated some patients.

      2. J says:

        Not sure about these two trials but there has been a published clinical trial since about a year, in STM.

  2. Emjeff says:

    I am surprised anyone thinks this is a worthwhile path to investigate. Yes, the BBB is good at keeping things out, but there are plenty of drugs that make it through. This is using a sledgehammer to kill a fly on your glass-top coffee table. Silly.

    1. David says:

      Depends how big the fly is, maybe?
      If you care more about killing the fly than the table, and you’re really, really careful with the hammer…

    2. NJBiologist says:

      Yeah, I’m with you on this one, Emjeff. Arterial gas embolism seems like an awfully drastic way to attempt drug delivery… seems like it would be way too effective at activating complement, for example.

  3. Mike says:

    What are the advantages of using techniques like this versus just administering a drug into the CSF? I know there are risks with a CSF port, but it seems like problem that we already have some solutions for. If a drug works, I wouldn’t let the delivery method stop people from using it, even if the route of administration isn’t ideal.

    1. Ex Expat says:

      For one thing, the site of BBB opening can be controlled precisely rather than increasing drug delivery throughout the brain (and possibly throughout the body). In addition, there may be beneficial effects in Alzheimer’s patients just from the bubbles + ultrasound combination in the absence of added drugs (link:

      1. Lane Simonian says:

        Even if ultrasound can be used to activate microglia to remove amyloid plaques it probably does not make any difference as amyloid plaques are not the probable cause of Alzheimer’s disease.

        The more likely benefit in mice at least is the activation of the phosphatidylinositol 3-kinase/Akt pathway. This pathway is critical for neurogenesis, synaptogenesis, proper blood flow, and the transport of glucose in the hippocampus. Unfortunately in Alzheimer’s disease this pathway is largely blocked by nitration.

        Certain anitoxidants are likely also involved in de-nitration. Of these methoxyphenols are probably the category of antioxidants most suited for the treatment of Alzheimer’s disease. Curcumin is one methoxphenol but it does not enter the bloodstream well. Another methoxyphenol is eugenol which also does not enter the bloodstream well but when inhaled has very little blood-brain barrier to cross to get to the brain.

        Yet another methoxyphenol is ferulic acid. HIgh concentrations of this methoxyphenol are needed because the compound does not enter into the fatty tissue of the brain very well. Heat processed ginseng contains high concentrations of ferulic acid and another methoxyphenol syringic acid. Here are the results from a small-scale clinical trial.

        The treatment groups showed significant improvement on the MMSE and ADAS. Patients with higher dose group (4.5 g/day) showed improvements in ADAS cognitive, ADAS non-cognitive, and MMSE score as early as at 12 weeks, which sustained for 24-week follow-up.
        These results demonstrate the potential efficacy of a heat-processed form of ginseng on cognitive function and behavioral symptoms in patients with moderately severe AD.

        1. David says:

          Is there a non-paywalled site that reports on this? You don’t list a place to read the study.

          1. Lane Simonian says:

            Here is a link to the abstract:


            I was able to get the full article through interlibrary loan, but otherwise there probably is a charge for access.

    2. Ben says:

      The main problem with using CSF for delivery is the very simple one that it doesn’t really work.

      1. pete says:

        ..but more recent findings that there’s CNS ==> lymphatic drainage at the dural sinuses ( reviewed here: ) suggests that maybe there’s a CNS / drug delivery angle that’s waiting to be exploited..?

      2. NJBiologist says:

        Ben, I’d be really careful about basing any important decisions on that paper. The author is a big name, but he makes a lot of assertions without obvious support. Based on his reasoning, ICV peptide delivery in rats should never work. (Based on both literature and my experience, I can tell you it does, quite well, for multiple peptide receptors, at doses far below anything effective IV.)

        1. Ben says:

          thanks, can you recommend a more balanced review in that case?

        2. Dr CNS says:

          For cases where the receptor you are targeting is localized within a few (I recall a 2-3 mm figure for rodent) mm from the ventricle, your ICV dosed peptide may reach it by diffusion.
          Otherwise, it is less likely.

          1. Dr CNS says:

            … of course, if the target is in the circumventricular organs, where the BBB is less tight, peptides will get to it too from blood…

    3. M says:

      Quite simple – the drug from CSF doesn’t easily enter the brain, and drug levels in CSF are not reflective of concentrations in the brain. The arachnoid barrier is partly responsible for this.
      Someone mentioned this is like a sledgehammer on a fly – I would argue that hyperosmotic mannitol is more in that category – patients need GA before administration.
      While med chem is argued above to ‘save the day’ and just get drugs into the brain, 50 years of GBM failure would argue against that point.

  4. JB says:

    Doesn’t mannitol already open up the BBB? Why do you need bubbles and ultrasound? Is it supposed to be better than mannitol?

  5. SA says:

    Hi Derek, did you hear about this news item:

    I would love to hear your and other readers view.

    1. Derek Lowe says:

      Yep, there’s a post coming on this stuff. You can probably imagine my overall reaction, though!

    2. Morten G says:

      I saw some very interesting comments on Reddit about this. I think in r/ELI5. You might want to have a look there.

  6. Barry says:

    Is it OK if I shake the infant selectively?

  7. BernYeeIris says:

    Ow! My Corpus Collusum!

  8. enotty says:

    The idea of a mechanical violation of the blood-brain barrier is fraught. Color me skeptical.

  9. Good PI for once says:

    Great reminder that biology is not done in a 1000 years. Its done in 10. Was relativity discovered by a 1000 men? No, one man, over 50 or so years. Ultrasound was a ” great technique “. Its not now. Wait until you know something before you claim knowlege. Or, as the ultrasound case shows here, you hurt kids…kids for a fricking nih grant you hurt kids????

    1. Big Man says:

      Well, good reminder that dumb feminist scientists that think that they have to monitor every facet of the reproductive cycle ( ie. Whole Life ) are misguided. For example, ultrasound kills your kid. Another example, being selfish is not going to help your kid.

  10. tangent says:

    Quasi-relatedly — does anyone have speculation what type of “sonic attack” might be involved in harm to diplomats in Cuba? Including traumatic brain injury symptoms, reportedly, in addition to hearing loss and inner-ear problems.

    A doctor who evaluated American and Canadian diplomats working in Cuba diagnosed them with conditions as serious as mild traumatic brain injury and damage to the central nervous system, CBS News said on Wednesday, citing medical records it reviewed.

    Officials are investigating whether the diplomats were targets of some form of sonic attack directed at their homes

  11. MoMo says:

    MIcrowaves. You can burn through a cranium at 300 yards with a focused beam. They are lucky. These devices were developed years ago and everyone has them.

    1. tangent says:

      For the Cuba thing? Sure but burns aren’t the symptom, nor the cataracts that are the more sensitive detector of microwave exposure.

  12. Mike says:

    Hi Derek,
    CarThera has published results of a clinical study using this technique in Science Translational Medicine last year: The safety of the technique has been shown in more than 15 patients to date. The study protocol is also available on

  13. Tao says:

    Hi Derek,

    We published a paper last week on this topic in PNAS. A real-time controller was designed and validated to open the BBB in a more precise and safer manner. Check this out here:


Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.