Skip to main content
  1. Courses
  2. NCS Podcast Series
  3. Episode 127: MASTER CLASS -...

Episode 127: MASTER CLASS - Non-Invasive Intracranial Pressure Monitoring

Loading audio, please wait...

Description

In this week's Master Class, Dr Stephan Mayer and Dr Jon Rosenberg are joined by Dr Krishna Rajajee to discuss non-invasive intracranial pressure monitoring

Contributors

  • Krishna Rajajee, MD

    Dr. Venkatakrishna Rajajee completed his medical training at the Madras Medical College in India followed by residency training in Neurology at the Mount Sinai Medical Center, New York in 2003, and fellowships in Stroke/ Vascular Neurology and Neurological Critical Care at the UCLA Medical Center 2003-2005. In addition, he has several years of experience in private practice as an attending physician in medical/ surgical intensive care units. He joined the University of Michigan’s neurosurgery department 2008 as faculty neurointensivist and vascular neurologist. He has served as the medical director of neurocritical care since 2014 and as co-director of the comprehensive stroke center since 2017.

    Dr. Rajajee has had a clinical, educational and research interest in critical care ultrasound since 2007. He successfully completed the American College of Chest Physicians (ACCP) critical care ultrasound program 2012-2013 and is a testamur of the National Board of Echocardiography. He has published several articles on optic nerve ultrasound and transcranial doppler in neurocritical care, and helped pioneer the use of real-time ultrasound guidance during bedside tracheostomy in 2011. Dr. Rajajee developed and directed the Neurocritical Care Society’s (NCS) annual critical care ultrasound workshop between 2012-2014, and has served on the Society of Critical Care Medicine’s (SCCM) Ultrasound Committee. He currently co-ordinates Michigan Medicine’s critical care ultrasound curriculum.
    His other interests include the monitoring and management of patients with severe Traumatic Brain Injury (TBI). In 2013, Michigan Medicine signed a memorandum of understanding on collaboration in research and education with the All India Institute of Medical Sciences (AIIMS), New Delhi, with TBI as a major focus. As part of this collaboration, Dr. Rajajee has received funding from the National Institutes of Health (NIH) to study techniques of intracranial pressure assessment applicable to low and middle income nations. He is also a co-investigator on a study funded by the Department of Defense (DoD) evaluating noninvasive techniques to assess the brain following TBI. He is the author of UpToDate’s topic reviews on the management and epidemiology of TBI.

    Dr. Rajajee has served on several national committees, including the NCS annual meeting committee, NCS physician accreditation certification and training (PACT) committee and the SCCM multidisciplinary critical care knowledge assessment program (MCCKAP) committee. He was co-chair of the Emergency Neurological Life Support (ENLS) airway, ventilation and sedation protocol subcommittee 2016-2020. In recognition of his contributions, Dr. Rajajee received a presidential citation from the neurocritical care society in 2013 & 2014.
    He is currently co-chair elect of the Neurocritical Care Society’s Guidelines Committee.

  • Stephan A. Mayer, MD, FCCM

    William T. Gossett Endowed Chair of Neurology
    Henry Ford Health System
    Stephan Mayer is Chair of Neurology for the Henry Ford Health System in Detroit. He previously served as director of Neurocritical Care for the Mount Sinai Health System in New York and director of the Neurological Intensive Care Unit at Columbia University College of Physicians and Surgeons in New York. Dr Mayer earned his medical degree from Cornell University Medical College. He completed a residency in neurology and a fellowship in critical care neurology at the Neurological Institute of New York, Columbia-Presbyterian Medical Center. He is board certified in neurology and a founding member and past president of the Neurocritical Care Society. Dr Mayer has published more than 230 original research articles, 180 review articles, and 340 abstracts, and he has written 7 books, and co-edited the most recent edition of Merritt’s Textbook of Neurology. He was principal investigator of the FAST Trial, a worldwide multicenter clinical trial evaluating ultra-early hemostatic therapy for brain hemorrhage, and he served as principal investigator of the NIH-funded New York Presbyterian Hospital hub of the Neurological Emergencies Treatment Trials (NETT) network. His work in helping victims of severe brain injury has been featured in The Wall Street Journal and the book Cheating Death by CNN medical correspondent Dr Sanjay Gupta.

  • Jon Rosenberg, MD

    Jon Rosenberg is an Assistant Professor of Neurology and Neurosurgery at Westchester Medical Center New York Medical College and Associate Program Director of the Neurocritical Care Fellowship at Westchester Medical Center. 

  1. 00:00:20
    Hello. This is John Rosenberg. Welcome
  2. 00:00:24
    back to the NCS podcast. We're coming
  3. 00:00:26
    to you from Westchester Medical Center.
  4. 00:00:27
    I'm joined with Dr. Stefan Mayer and Dr.
  5. 00:00:30
    Krishna Rajaji. How are you doing
  6. 00:00:33
    Krishna?
  7. 00:00:36
    Good to see you.
  8. 00:00:40
    This is the master class and we're going
  9. 00:00:43
    to be talking today about non-invasive
  10. 00:00:48
    intracranial pressure monitoring. We
  11. 00:00:50
    have an expert here who's funded
  12. 00:00:53
    research. I was focused on non-invasive
  13. 00:00:56
    SCP monitoring and today we're going to
  14. 00:00:59
    pick your brain. Dr. Rajaji for our
  15. 00:01:00
    listeners who maybe haven't heard of you
  16. 00:01:01
    already. Dr. Rajaji is
  17. 00:01:07
    a neurointensivist and vascular
  18. 00:01:08
    neurologist, professor of neurology in
  19. 00:01:10
    neurosurgery at Michigan where he serves
  20. 00:01:12
    as the medical director of the
  21. 00:01:13
    neurocritical care and co-director of
  22. 00:01:14
    their comprehensive stroke center. Dr.
  23. 00:01:17
    Rajaji is a pioneer in critical care
  24. 00:01:19
    ultrasound and non-invasive ICP
  25. 00:01:20
    monitoring. He's also the former chair
  26. 00:01:22
    of the NCS Guidelines Committee. I
  27. 00:01:25
    encourage all the listeners to review
  28. 00:01:26
    his article non-invasive intracranial
  29. 00:01:29
    pressure monitoring. Are we there yet?
  30. 00:01:31
    Published in neurocritical care last
  31. 00:01:33
    month. In addition to his prior NCS
  32. 00:01:35
    podcast on Optognosheet diameter and the
  33. 00:01:37
    NCS guideline review. Dr. Rajaji,
  34. 00:01:39
    welcome back to the podcast. I hope the
  35. 00:01:41
    intro does you justice. And I think
  36. 00:01:44
    it's fair to say, you're a friend of
  37. 00:01:45
    the show now I am, yeah, no, it's
  38. 00:01:47
    been great being on the podcast. Thank
  39. 00:01:49
    you so much for having me back to. Sure,
  40. 00:01:52
    so let's get into, let's talk a little
  41. 00:01:54
    bit about non-vasive intracranial
  42. 00:01:57
    pressure monitoring. I think we have
  43. 00:02:00
    TCDs, peopleometry, optognosheet
  44. 00:02:02
    diameter. Maybe let's start with one of
  45. 00:02:06
    those and we can review all three and
  46. 00:02:08
    tell us a little bit about your thoughts
  47. 00:02:10
    on them and how you use them in practice.
  48. 00:02:14
    Sure, so I think maybe started off a
  49. 00:02:16
    nursery diameter then? Perfect. Yeah.
  50. 00:02:21
    So, you know, among the various
  51. 00:02:22
    non-invasive means that we use to
  52. 00:02:24
    measure intranial pressure, I think
  53. 00:02:27
    optic nursery diameter measurement is
  54. 00:02:28
    probably the best studied, arguably the
  55. 00:02:32
    best studied, sort of originates the
  56. 00:02:34
    concept of papilladema. So, you know,
  57. 00:02:36
    some people like to refer to it as
  58. 00:02:38
    papilladema 20. And a lot of the early
  59. 00:02:43
    work was done using autopsy specimens by
  60. 00:02:47
    Helmke and Hansen in Germany, where
  61. 00:02:50
    they're basically infused, blew it into
  62. 00:02:52
    the sub-brachnoid space, you know,
  63. 00:02:55
    within the optic nerve sheet, and
  64. 00:02:57
    demonstrated that when you simulate an
  65. 00:02:59
    increase in ICP, the optic nerve sheet
  66. 00:03:01
    distends, as you would expect it to,
  67. 00:03:03
    and that you can measure that using
  68. 00:03:04
    ultrasound. So, you know, eventually
  69. 00:03:06
    that's followed by a bunch of human
  70. 00:03:08
    studies where you perform an ultrasound
  71. 00:03:12
    of the eyeball, you measure the
  72. 00:03:15
    diameter of the optic nerve sheet. at a
  73. 00:03:17
    specific point, at a fixed point behind
  74. 00:03:20
    the eye, because depending on how far
  75. 00:03:22
    back behind the eye, you measure it,
  76. 00:03:24
    you're gonna get very different
  77. 00:03:25
    measurements. But, so you use the same
  78. 00:03:28
    point, same distance each time, and
  79. 00:03:31
    you measure the optic nursery diameter,
  80. 00:03:32
    and then correlate it to simultaneous in
  81. 00:03:34
    ways of internal pressure monitoring.
  82. 00:03:37
    And so sort of, there's been a little
  83. 00:03:40
    bit of a saga with ONSD measurement. I
  84. 00:03:42
    don't think the story's ended yet, but
  85. 00:03:43
    I think that with, like with a lot of
  86. 00:03:46
    other critical care ultrasound tools,
  87. 00:03:49
    sort of the early spay to studies were
  88. 00:03:51
    very positive. We did one of those
  89. 00:03:53
    studies and, you know, we saw pretty
  90. 00:03:56
    good correlation within ways of ICP. I
  91. 00:04:01
    have since transitioned to being a
  92. 00:04:03
    little bit more of a skeptic and I think
  93. 00:04:06
    gone from being one of the proponents to
  94. 00:04:08
    being a little bit more skeptical about
  95. 00:04:10
    at least clinical use. It's certainly,
  96. 00:04:12
    it's not like, you know, there hasn't
  97. 00:04:14
    been sufficient time to evaluate this
  98. 00:04:15
    tool. has. We've spent well over a
  99. 00:04:19
    decade now studying this tool. And at
  100. 00:04:22
    least based on some of the work that
  101. 00:04:24
    we've done, where we attempted to
  102. 00:04:27
    address some of the sources of bias in
  103. 00:04:30
    these studies. Bias in these studies is
  104. 00:04:32
    a big problem. So addressing them, I
  105. 00:04:34
    think, is really important.
  106. 00:04:37
    At least one of our studies that we try
  107. 00:04:38
    to address bias results were not nearly
  108. 00:04:41
    as good as we had hoped. And I can talk
  109. 00:04:44
    more about the reasons that I think that
  110. 00:04:45
    we should be careful about using this
  111. 00:04:47
    day-to-day in clinical practice. But I
  112. 00:04:51
    think that I think the bottom line is,
  113. 00:04:55
    there's clearly a biological
  114. 00:04:57
    relationship, as you would expect.
  115. 00:04:59
    Just as with Papalirima, there's
  116. 00:05:00
    clearly a biological relationship
  117. 00:05:02
    between high ICP and optic nerve sheet
  118. 00:05:04
    distension, but in terms of being able
  119. 00:05:07
    to use it at the bedside in ICU to
  120. 00:05:09
    decide when to initiate treatment of ICP,
  121. 00:05:11
    when to not initiate treatment of
  122. 00:05:13
    elevated ICP, that I think is more.
  123. 00:05:16
    problematic. Let me ask a question and
  124. 00:05:20
    what about using it in your clinical
  125. 00:05:22
    practice to help you decide when to put
  126. 00:05:25
    a bolt in. Your review article starts
  127. 00:05:29
    with a scenario of a full-man hepatic
  128. 00:05:32
    failure patient and you know an elevated
  129. 00:05:35
    INR.
  130. 00:05:37
    You know you can reverse the
  131. 00:05:39
    anticoagulation and put it in. There's
  132. 00:05:42
    still some residual risk it's probably
  133. 00:05:44
    higher than a non-pregal pathic patient
  134. 00:05:46
    but do you ever do that to select
  135. 00:05:49
    patients for invasive monitoring? So
  136. 00:05:53
    we've never used so suggest to give you
  137. 00:05:55
    the background of that Stefan. So we
  138. 00:05:58
    the University of Michigan we had a we
  139. 00:06:00
    have a long-standing protocol for
  140. 00:06:02
    invasive monitoring in patients with
  141. 00:06:04
    acute lower failure. We put the
  142. 00:06:06
    protocol in place on 2011 and you know
  143. 00:06:10
    we've had over a decade of using that
  144. 00:06:13
    protocol. the fact that we regularly
  145. 00:06:15
    did invasive monitoring these patients
  146. 00:06:17
    allowed us to then study non-invasive
  147. 00:06:21
    techniques more accurately. So there
  148. 00:06:24
    were several years in which we
  149. 00:06:25
    simultaneously performed ONSD
  150. 00:06:27
    measurement, traskin, or doppler along
  151. 00:06:29
    with ICP measurement and then correlated
  152. 00:06:32
    against the gold standard. ONSD did not
  153. 00:06:35
    perform so well. So we published our
  154. 00:06:39
    experience at invasive monitoring after
  155. 00:06:42
    the first five years of using the
  156. 00:06:44
    protocol. We followed that up with
  157. 00:06:47
    another publication of the article here,
  158. 00:06:49
    where we compared our non-invasive
  159. 00:06:52
    measurements with against the gold
  160. 00:06:54
    standard of invasive. ONSD really
  161. 00:06:56
    didn't do well at all. To the extent
  162. 00:06:59
    that I would not use it, we also looked
  163. 00:07:02
    at transcranial doppler in this context,
  164. 00:07:05
    and transcranial doppler did better. I
  165. 00:07:06
    mean, it's by no means a perfect tool.
  166. 00:07:09
    It's an extremely flawed tool. but it
  167. 00:07:11
    actually did better than our technology
  168. 00:07:13
    diameter in our liver failure patients.
  169. 00:07:16
    So I mean, so to answer your question,
  170. 00:07:18
    I wouldn't, I think I would use other
  171. 00:07:22
    tools to select patients for invasive
  172. 00:07:25
    monitoring, if you were gonna do it at
  173. 00:07:26
    all. I'm happy to discuss sort of an
  174. 00:07:28
    approach to your approach versus our
  175. 00:07:30
    approach to acute lower failure because
  176. 00:07:33
    it's, you know, we all know there's no
  177. 00:07:34
    one, you know, perfect approach to
  178. 00:07:36
    these patients, but I would not use ONS
  179. 00:07:39
    to you to select patients. I think sort
  180. 00:07:42
    of stuff is what we've come to over the
  181. 00:07:44
    years. And it seems like for our
  182. 00:07:48
    listeners, one of the reasons, and
  183. 00:07:50
    correct me if I'm wrong, some of the
  184. 00:07:51
    reasons are one, the interrelated
  185. 00:07:54
    reliability seems to be quite poor
  186. 00:07:56
    unless you follow up training, so
  187. 00:07:58
    different examiners. We see in real
  188. 00:08:01
    world with different examiners doing it,
  189. 00:08:03
    you can get widely different
  190. 00:08:04
    measurements in the same, you know, on
  191. 00:08:07
    the same patient and then more to the
  192. 00:08:09
    end. different studies have shown
  193. 00:08:11
    different cutoffs. So depending on what
  194. 00:08:13
    you pick as your cut off, if you want a
  195. 00:08:15
    distention greater than 5 millimeters,
  196. 00:08:17
    55, 7, you may have a different
  197. 00:08:19
    sensitivity, specificity, not those
  198. 00:08:22
    sensitivities and specificities actually
  199. 00:08:23
    haven't worn out to be the same at
  200. 00:08:25
    different cutoffs.
  201. 00:08:27
    Absolutely. Yeah, now you hit the nail
  202. 00:08:29
    on the head. I mean, also very much
  203. 00:08:31
    the two issues. I think that if you've
  204. 00:08:33
    got people who are experienced with the
  205. 00:08:36
    measurement of ONSD, just taking a step
  206. 00:08:38
    back, there's actually debate about,
  207. 00:08:40
    you know, how best to measure ONSD at
  208. 00:08:42
    all among experts. But let's assume
  209. 00:08:45
    that you all agree on one way to do it,
  210. 00:08:47
    and then you are experienced in doing it
  211. 00:08:49
    in that way, and you teach others, and
  212. 00:08:51
    they do it the same way. Over time,
  213. 00:08:53
    you can get in a very good interview to
  214. 00:08:55
    the viability. But that's not the
  215. 00:08:57
    reality of clinical practice. The
  216. 00:08:59
    reality of clinical practice is that,
  217. 00:09:00
    you know, a person of the bedside doing
  218. 00:09:02
    this is often not experienced. There
  219. 00:09:03
    are often fellows, you know, residents,
  220. 00:09:07
    and, you know, maybe they've done it a
  221. 00:09:09
    few things. maybe you've shown it to
  222. 00:09:10
    them a few times, and then the
  223. 00:09:12
    interrelated reliability really may not
  224. 00:09:14
    be as good as you think it is. I mean,
  225. 00:09:16
    that's what we found in our study,
  226. 00:09:17
    where we went through this process of
  227. 00:09:20
    training, you know, folks at the
  228. 00:09:21
    bedside, the providers at the bedside,
  229. 00:09:24
    to do this, we had to check them off.
  230. 00:09:27
    So we had the experienced people, you
  231. 00:09:30
    know, had them do with 20 studies and
  232. 00:09:32
    then check them off. But then after we
  233. 00:09:36
    did 50 studies, we saw that there was
  234. 00:09:38
    very poor correlation between export
  235. 00:09:40
    measurements and the inexperienced folks.
  236. 00:09:42
    So we went back and cleaned all of these
  237. 00:09:44
    folks again. And then you had really
  238. 00:09:46
    good correlation. But you know, you've
  239. 00:09:49
    got to be very cognizant of the fact
  240. 00:09:51
    that just because you trained people to
  241. 00:09:53
    do this, it doesn't necessarily mean
  242. 00:09:55
    that they're going to do it exactly like
  243. 00:09:57
    you would want them to. And this is
  244. 00:09:59
    certainly not a unique topic now, she
  245. 00:10:01
    diameter measurement. We've seen is the
  246. 00:10:03
    exact same thing with other forms of
  247. 00:10:05
    critical canal to solve.
  248. 00:10:12
    Why don't we jump to TCBs? Maybe we
  249. 00:10:13
    could talk a little bit about your
  250. 00:10:14
    experience with TCTs, the literature
  251. 00:10:17
    behind them, and how you use them in
  252. 00:10:19
    clinical practice. For sure. So, I
  253. 00:10:22
    think the
  254. 00:10:30
    thing to start with is that there's no
  255. 00:10:32
    single way to do this. We focus on the
  256. 00:10:35
    sosnika technique, but there's actually
  257. 00:10:37
    many different ways that you can use
  258. 00:10:38
    transcranial to apply to SSICP, ranging
  259. 00:10:42
    from relatively simple, your
  260. 00:10:44
    possibility index to extremely the that
  261. 00:10:46
    think I But. complicated
  262. 00:10:50
    best evaluated way to do this is using
  263. 00:10:53
    the sosnika technique. Myexosnika
  264. 00:10:57
    developed this technique at Cambridge
  265. 00:10:58
    back in the late '90s. And essentially,
  266. 00:11:01
    what
  267. 00:11:05
    you do is you measure four velocities in
  268. 00:11:07
    the middle cerebral arteries
  269. 00:11:07
    simultaneously measure the. mean
  270. 00:11:08
    arterial pressure from an arterial
  271. 00:11:10
    catheter and you use a formula that he
  272. 00:11:12
    developed to estimate the cerebral
  273. 00:11:14
    perfusion pressure. You mean,
  274. 00:11:16
    obviously you know the mean arterial
  275. 00:11:18
    pressure and so you calculate the
  276. 00:11:20
    intracranial pressure. So over the
  277. 00:11:22
    years that's been best evaluated, to my
  278. 00:11:24
    knowledge, it's one of the few tools
  279. 00:11:26
    that's been evaluated in a pretty high
  280. 00:11:30
    quality study, I would say, sort of
  281. 00:11:32
    it's an international multi-center
  282. 00:11:34
    prospective study was impressive too And
  283. 00:11:38
    so, I mean, that's really what we use,
  284. 00:11:41
    that's what we've evaluated in all our
  285. 00:11:43
    studies, liver failure, and sort of
  286. 00:11:45
    general neurocritical care populations.
  287. 00:11:49
    There's, the thing about TCD is, it's
  288. 00:11:53
    harder to do really than aprychnosis
  289. 00:11:55
    diameter in some ways, I mean, you
  290. 00:11:57
    really probably should be certified in
  291. 00:12:00
    TCD. You probably should be able to do
  292. 00:12:03
    it well. And then there's a bunch of
  293. 00:12:06
    technical considerations. You know,
  294. 00:12:08
    for example, our recent studies
  295. 00:12:11
    suggested that you should perform angle
  296. 00:12:13
    correction, which you can only do if
  297. 00:12:15
    you've got duplex, transcranial color
  298. 00:12:17
    code, it's anography. If you've got an
  299. 00:12:19
    regular TCD machine, you probably can't
  300. 00:12:22
    do that. I think there's the where you
  301. 00:12:25
    level the transducer, whether you do it
  302. 00:12:27
    at the plebostatic axis of the plagueus
  303. 00:12:29
    has probably a pretty significant impact.
  304. 00:12:32
    So there are all these technical
  305. 00:12:33
    considerations Now, if you're able to
  306. 00:12:36
    do it, then I think our experience is
  307. 00:12:39
    that it's like a lot of other
  308. 00:12:41
    noninvasive tools, it's probably best
  309. 00:12:45
    used as a rule out tool in that if it
  310. 00:12:47
    looks entirely normal, if your
  311. 00:12:51
    calculated ICP is normal, it's not
  312. 00:12:55
    perfect, but there's a decent chance
  313. 00:12:57
    that the ICP is actually normal If you
  314. 00:13:00
    calculate ICP using the technique that
  315. 00:13:03
    we do, and I just want to be very
  316. 00:13:04
    particular about that, using the
  317. 00:13:06
    technique we do. which is leveling to
  318. 00:13:08
    the Traegas, performing angle
  319. 00:13:10
    correction. Using technique, we do
  320. 00:13:13
    have the calculated ICPs elevated,
  321. 00:13:15
    really doesn't mean that much, unless
  322. 00:13:17
    it's extremely elevated. And so in that
  323. 00:13:20
    case, you just need to look at your
  324. 00:13:22
    other modalities, maybe poplometry,
  325. 00:13:25
    maybe something else, and just to see
  326. 00:13:27
    your measurements, to see where things
  327. 00:13:29
    are at. Krishna, why don't you also
  328. 00:13:34
    look at the straight up Positility Index
  329. 00:13:36
    at the same time, and you've got to
  330. 00:13:37
    adopt a probe there. Is there any way
  331. 00:13:39
    to integrate the two techniques to maybe
  332. 00:13:42
    improve the sensitivity?
  333. 00:13:46
    Yeah, so I mean, so to answer the
  334. 00:13:48
    first question, Stefan, yes, you can
  335. 00:13:50
    look at the Positility Index. Our study
  336. 00:13:54
    and other studies, I think, suggest
  337. 00:13:55
    that it's not as robust a tool as
  338. 00:13:59
    calculating the ICP, but it's much
  339. 00:14:01
    easier to do, you know, for sure It
  340. 00:14:03
    doesn't require someone who knows the
  341. 00:14:06
    specific degree. technique, you can
  342. 00:14:07
    get a tech to perform the TCD and, you
  343. 00:14:10
    know, you just give it get a positive
  344. 00:14:11
    utility index. So, the positive index
  345. 00:14:15
    that is entirely normal, you know,
  346. 00:14:18
    that is less than I would say, you know,
  347. 00:14:21
    08 is
  348. 00:14:24
    relatively reassuring. It's relatively
  349. 00:14:26
    reassuring. It's not entirely
  350. 00:14:28
    reassuring though. So, you know,
  351. 00:14:29
    whatever TCD technique that you use, I
  352. 00:14:33
    think to understand is, if you just
  353. 00:14:35
    look at all of the diagnostic accuracy
  354. 00:14:36
    studies, all of these studies have a
  355. 00:14:38
    pretty large confidence interval. None
  356. 00:14:41
    of these studies really have enough
  357. 00:14:42
    patience to have a tight, tight
  358. 00:14:45
    confidence interval. So, there's a
  359. 00:14:47
    pretty significant room for error. And
  360. 00:14:49
    if you look at the best study of this
  361. 00:14:51
    technique, which is impressive too,
  362. 00:14:55
    then, you know, they focus on the
  363. 00:14:56
    negative predictive value, which is 96.
  364. 00:14:59
    But the negative predictive value is not
  365. 00:15:01
    necessarily the right number to look at.
  366. 00:15:03
    It's probably the sensitivity that you
  367. 00:15:05
    want to look at. And the sensitivity is
  368. 00:15:08
    really not that good. It's maybe, I
  369. 00:15:10
    think, the 80s. So you can use it,
  370. 00:15:15
    whether you're using the estimated ICP
  371. 00:15:18
    as a positive index, you can use it.
  372. 00:15:19
    You can get some relatively assurance
  373. 00:15:22
    from a normal value, but I wouldn't run
  374. 00:15:25
    with any one modality alone. So that
  375. 00:15:28
    leads to the second part, if you're a
  376. 00:15:29
    question, what if you integrated it
  377. 00:15:31
    with other numbers? I think that's
  378. 00:15:33
    basically what we do If the PI is
  379. 00:15:36
    entirely normal, the estimated ICP is
  380. 00:15:38
    probably going to be close to normal
  381. 00:15:40
    because the principle is a lot similar.
  382. 00:15:45
    But I would not integrate it with
  383. 00:15:49
    another TCD modality. I would integrate
  384. 00:15:51
    it with entirely different modalities,
  385. 00:15:53
    whether it's pupalometry or CT scan,
  386. 00:15:58
    whatever else you have, to be honest.
  387. 00:16:00
    Because one thing I think we have
  388. 00:16:03
    learned is, Highly excessively on one
  389. 00:16:07
    tool or one measurement that can really
  390. 00:16:10
    lead you, you know, clung.
  391. 00:16:14
    I can summarize what you just said as
  392. 00:16:16
    well. There was a lot of high level
  393. 00:16:19
    critical care and ultrasound just
  394. 00:16:22
    described in that SNP, but basically
  395. 00:16:24
    what you're saying, so TCDs in the most
  396. 00:16:26
    validated form, which is the formula
  397. 00:16:28
    with Chosnica with continuous TCDs and a
  398. 00:16:31
    formula to measure ICP, CPP based on
  399. 00:16:35
    your flow velocities, that's been the
  400. 00:16:38
    most validated. In doing that, you
  401. 00:16:41
    found one,
  402. 00:16:44
    you actually, in the impressive trial,
  403. 00:16:45
    and when most of the literature has
  404. 00:16:46
    shown is that the strength of that,
  405. 00:16:48
    even in our best modality, is the
  406. 00:16:51
    negative predictive value, that there's
  407. 00:16:52
    a high negative predictive value, but
  408. 00:16:54
    the positive predictive value for
  409. 00:16:55
    elevated for TCDs and the sensitivity is
  410. 00:16:58
    not great, which makes it difficult to
  411. 00:17:01
    be solely to rely on it as a screening
  412. 00:17:03
    test.
  413. 00:17:05
    The other point that you make is that
  414. 00:17:06
    TCD, so when you dokyl
  415. 00:17:09
    ultrasoundography is very angle
  416. 00:17:10
    dependent. So if you're, if you're
  417. 00:17:13
    insulating the velocity of the MCA and
  418. 00:17:15
    you don't have the correct angle of your
  419. 00:17:16
    angles off by over 20 degrees, then
  420. 00:17:19
    your flow velocity may be off. So it's
  421. 00:17:21
    very, again, operator dependent. And
  422. 00:17:23
    then you may not, you may be
  423. 00:17:25
    underestimating your true flow velocity.
  424. 00:17:28
    The other interesting points, it sounds
  425. 00:17:29
    like for when you do TCDs, you level
  426. 00:17:32
    the art line at the tragus, as opposed
  427. 00:17:34
    to the phlebostatic axis. Did I catch
  428. 00:17:36
    that? So we level at the phlebostatic
  429. 00:17:39
    axis, to be honest.
  430. 00:17:42
    So with implicit, so yeah. So we level
  431. 00:17:44
    at the phlebostatic axis. The reason we
  432. 00:17:46
    do that is that's been the culture at
  433. 00:17:49
    our ICU going back a very long time.
  434. 00:17:51
    It's very difficult to change that. So
  435. 00:17:53
    even when you conduct research, it's
  436. 00:17:56
    just, it's too much to try to tell
  437. 00:18:00
    people change the plants use their level
  438. 00:18:02
    maybe maybe we should but we we We just
  439. 00:18:05
    found it more convenient, I guess, to
  440. 00:18:08
    go with the standard clinical practice
  441. 00:18:09
    on our unit. Impressive two, though,
  442. 00:18:12
    they did level to the
  443. 00:18:15
    Tragus. And you know, Impressive two
  444. 00:18:16
    is an international multi-center
  445. 00:18:18
    prospective study, so if you want to go
  446. 00:18:21
    with the best violated technique, it's
  447. 00:18:24
    probably their technique, and they
  448. 00:18:26
    level to do the Tragus. But they, and
  449. 00:18:29
    that also points to the overall, one of
  450. 00:18:32
    the issues with not a days in monitoring
  451. 00:18:33
    is if you're gonna use the technique
  452. 00:18:34
    then just like a form if you don't level
  453. 00:18:37
    the A line at the Tragus, then
  454. 00:18:39
    technically your numbers aren't gonna be,
  455. 00:18:40
    your cut-offs aren't gonna be the same,
  456. 00:18:42
    your sensitivity is best to visit, but
  457. 00:18:43
    that cut-off won't be the same. Yeah,
  458. 00:18:45
    I think you just have to understand the
  459. 00:18:47
    impact of where you level it, because
  460. 00:18:49
    my understanding is that Chuznika
  461. 00:18:51
    actually leveled to the phlebostatic
  462. 00:18:52
    axis as we do.
  463. 00:18:56
    Impressive to, you know, their level
  464. 00:18:57
    to the Tragus. So it's more a question
  465. 00:19:00
    of understanding what the implications
  466. 00:19:02
    are. What we've seen, for example, is
  467. 00:19:05
    that if you level to the plebostatic
  468. 00:19:08
    axis, you see what we saw in our study,
  469. 00:19:11
    which is a consistent overestimation of
  470. 00:19:13
    ICP. So you just have to understand
  471. 00:19:16
    that if you do it using our technique,
  472. 00:19:17
    you're probably your lower estimate ICP.
  473. 00:19:20
    If you level to the tragus, from what
  474. 00:19:23
    my experience is, closer to the actual
  475. 00:19:26
    ICP, but then your sensitivity is
  476. 00:19:28
    actually lower.
  477. 00:19:31
    So yeah If I just want to, for our
  478. 00:19:35
    listeners, this is a very high level
  479. 00:19:37
    conversation and a lot of our listeners
  480. 00:19:40
    are learners. So just for point of
  481. 00:19:42
    clarification, I'm going to share my
  482. 00:19:44
    views about this whole leveling, your
  483. 00:19:47
    ICP. This may sound like an obscure
  484. 00:19:50
    discussion about, you know, none of
  485. 00:19:52
    these ICP, using transparently a
  486. 00:19:54
    Doppler with a mathematical formula that,
  487. 00:19:57
    you know, Soneka developed. This is a
  488. 00:19:59
    day-to-day issue in every nurse you run
  489. 00:20:02
    a world every day. when you transduce
  490. 00:20:05
    ICP off of a ventricular drain, right?
  491. 00:20:07
    Where do you level that pressure
  492. 00:20:09
    transducer? And my understanding of
  493. 00:20:13
    conventional wisdom is the true ICP is
  494. 00:20:16
    going to be measured in the case where
  495. 00:20:18
    you have the pressure transducer at the
  496. 00:20:22
    level of the tragus, which approximates
  497. 00:20:24
    the brominum and row. And that's really
  498. 00:20:27
    important when your head up, which many
  499. 00:20:29
    of our patients are. However, if you
  500. 00:20:32
    go head flat and you feel the patient
  501. 00:20:34
    can tolerate that, the level of the
  502. 00:20:36
    year or tragus isn't the level of the
  503. 00:20:38
    mid-exillary line, and it's kind of a
  504. 00:20:40
    non-issue. So I'll just, at this point,
  505. 00:20:44
    I'd like to ask you, Chris, to move
  506. 00:20:48
    over now to your views on pupalometry.
  507. 00:20:50
    Can you explain the rationale? Why
  508. 00:20:53
    would measuring pupal
  509. 00:20:56
    tell us about ICP? Stefan, can I
  510. 00:20:59
    squeeze in one more question before we
  511. 00:21:01
    do? Sure.
  512. 00:21:04
    Most of the setups in the ICUs, we do
  513. 00:21:06
    intermittent TCDs. The tech comes
  514. 00:21:08
    around those TCDs. Is this enough to
  515. 00:21:12
    get to use the child's neck of formula?
  516. 00:21:14
    Or do you really need continuous TCDs?
  517. 00:21:17
    Well, so it's, you know, what
  518. 00:21:19
    Chasnika did was not just continuous TCD,
  519. 00:21:21
    it was digital waveform integration,
  520. 00:21:25
    you know, with the waveforms from the
  521. 00:21:26
    monitor. So if you could do that with
  522. 00:21:29
    your intermittent TCD, I think that
  523. 00:21:31
    would be great I don't think it's really
  524. 00:21:33
    about continuous TCD as much as that
  525. 00:21:35
    integration of waveforms. So if you
  526. 00:21:38
    were to just stand there doing your
  527. 00:21:39
    intermittent TCD for, I don't know, 10
  528. 00:21:40
    minutes, five to 10 minutes, and then
  529. 00:21:43
    you have a system by which you can do a
  530. 00:21:45
    digital integration of waveforms.
  531. 00:21:48
    It's not that complicated, right? I
  532. 00:21:49
    mean, many of us do the PRX or
  533. 00:21:51
    something similar every day. If we had
  534. 00:21:54
    a tool like that, I think that would
  535. 00:21:55
    probably be technologically superior and
  536. 00:21:59
    probably better, whereas not the
  537. 00:22:00
    reality of what most of us have,
  538. 00:22:02
    available to us. So, most of us will
  539. 00:22:05
    probably continue to do just this
  540. 00:22:07
    intermittent TCD and then look at the
  541. 00:22:09
    screen and see what the meaner to your
  542. 00:22:10
    pressure is and have somebody in order
  543. 00:22:12
    So, okay.
  544. 00:22:19
    Why don't we move on to to people on the
  545. 00:22:21
    tree not tell us a little bit about
  546. 00:22:24
    about your youth about the literature on
  547. 00:22:25
    people on between how you use it. So,
  548. 00:22:28
    so, you know, I think I want to
  549. 00:22:29
    discussion like you guys have said,
  550. 00:22:31
    it's been a technical high level so far.
  551. 00:22:34
    People are at least probably the most
  552. 00:22:36
    intuitive. I think for most of us it's
  553. 00:22:39
    nature's ICP monitor And so, so it's
  554. 00:22:44
    honestly, you know, you asked me, I
  555. 00:22:46
    think that's it's great strength that it
  556. 00:22:48
    is easy to do and it's intuitive. We
  557. 00:22:51
    know that when your ICP is elevated,
  558. 00:22:55
    particularly when you're on the brink of
  559. 00:22:57
    cerebral herniation, then, you know,
  560. 00:22:59
    there's an impact on the parasympathetic
  561. 00:23:01
    tone transmitted through the, through
  562. 00:23:05
    the third nerve,
  563. 00:23:07
    and that's, you know, reflected in
  564. 00:23:11
    pupillary dilatation
  565. 00:23:13
    initially, because your pupillary
  566. 00:23:15
    fibers are on the outside.
  567. 00:23:18
    So, you know, that goes first and then
  568. 00:23:19
    eventually you're going to complete
  569. 00:23:20
    third up policy. And it's just that
  570. 00:23:23
    very simple principle that elevated ICP
  571. 00:23:26
    is going to impact your oculomotor nerve
  572. 00:23:29
    and therefore your pupils.
  573. 00:23:33
    So, oldest principle in your critical
  574. 00:23:34
    care, I think where pupalometry takes
  575. 00:23:38
    this a step further is that when you do
  576. 00:23:39
    this manually, it seems like none of us
  577. 00:23:44
    do it very well All of the data suggests
  578. 00:23:46
    that there's a high rate of error when
  579. 00:23:48
    you manually estimate sluggish pupils,
  580. 00:23:53
    even when you try to estimate pupillary
  581. 00:23:55
    size, non-reactivity, just even
  582. 00:23:57
    complete non-reactivity, there's just a
  583. 00:23:59
    high rate of error, there's several
  584. 00:24:01
    studies that have demonstrated that.
  585. 00:24:03
    Whereas pupalometry allows us to
  586. 00:24:06
    accurately quantify the pupillary light
  587. 00:24:09
    response
  588. 00:24:12
    and
  589. 00:24:14
    I think that's really where the value
  590. 00:24:16
    comes in.
  591. 00:24:19
    So, if you have sluggish pupils, if
  592. 00:24:22
    you have non-reactive pupils, in the
  593. 00:24:24
    right context, you want to worry about
  594. 00:24:26
    elevated ICP. There's additional value
  595. 00:24:30
    in that, you know, we know there's all
  596. 00:24:32
    these patients with temporal lobe
  597. 00:24:34
    hematomas. Maybe it's a big temporal
  598. 00:24:35
    lobe confusion, a traumatic brain
  599. 00:24:37
    injury, maybe it's some of the
  600. 00:24:38
    malignant MCS stroke, maybe it's a
  601. 00:24:40
    spontaneous intercerebral hemorrhage,
  602. 00:24:42
    but all of these patients might suffer
  603. 00:24:45
    on coordination because they have
  604. 00:24:47
    predominantly temporalopathology,
  605. 00:24:49
    even when the ICP is not particularly
  606. 00:24:51
    elevated and with a pupilometer, you
  607. 00:24:53
    would be able to detect that sooner. So
  608. 00:24:56
    I think the use of pupilometry to
  609. 00:24:58
    quantify the
  610. 00:25:01
    pupillary light response is very
  611. 00:25:04
    intuitive. There's companies that have
  612. 00:25:07
    made it easy for us by creating these
  613. 00:25:10
    proprietary indices the indices,
  614. 00:25:12
    integrate, you know, starting pupil
  615. 00:25:14
    size. the extent to the final size of
  616. 00:25:18
    the constriction, speed of constriction,
  617. 00:25:21
    speed of relaxation, and then create
  618. 00:25:22
    these indices which you can use to
  619. 00:25:26
    decide if your pupil is sluggish or
  620. 00:25:29
    normal. When you use this, I think to
  621. 00:25:33
    decide if your ICP is elevated or not,
  622. 00:25:36
    I think that while I'm certainly a fan
  623. 00:25:39
    be given how intuitive it is, it's
  624. 00:25:41
    important to note that data is actually
  625. 00:25:43
    not as robust It's just not being
  626. 00:25:46
    validated to nearly the same extent as
  627. 00:25:49
    even TCD or optic nerve sheet diameter.
  628. 00:25:53
    And so you just got to be careful about
  629. 00:25:54
    that. And when you do study it, it
  630. 00:25:57
    often turns out to be a little bit less
  631. 00:25:59
    accurate overall than for example, TCD.
  632. 00:26:03
    So you just got to be careful. It's so
  633. 00:26:04
    easy and convenient to use that I use it
  634. 00:26:09
    all the time. But this is yet another
  635. 00:26:11
    modality that I would just, I would not
  636. 00:26:14
    take that to the bank. in itself, I
  637. 00:26:18
    certainly would recognize that
  638. 00:26:23
    if you have a sluggish pupil, again,
  639. 00:26:26
    you want to pick that seriously, you
  640. 00:26:29
    want to evaluate that aggressively,
  641. 00:26:33
    but you should also recognize them, you
  642. 00:26:35
    don't have a sluggish pupil, your ICP
  643. 00:26:37
    could be elevated beyond a level that
  644. 00:26:39
    you would normally be treating it
  645. 00:26:44
    Yeah, just if I can share my views, I
  646. 00:26:49
    like to say, When should you start
  647. 00:26:51
    doing pupil on the tree? There's no
  648. 00:26:53
    going back. But I view
  649. 00:26:57
    at pupillary symmetries and
  650. 00:26:59
    abnormalities is more of an indication
  651. 00:27:02
    of brainstem distress from displacement
  652. 00:27:06
    and torsion, which is a natural
  653. 00:27:10
    consequence when ICP is both elevated
  654. 00:27:14
    and compartmentalized. And that's what
  655. 00:27:17
    exactly what you're going to get with
  656. 00:27:20
    tranestentorial herniation
  657. 00:27:27
    or self-hossing herniation. So it's
  658. 00:27:29
    kind of a, for me, anyway, a
  659. 00:27:29
    downstream sign that, you know, as
  660. 00:27:30
    long as there's some integrating a
  661. 00:27:31
    massive effect with the start of that
  662. 00:27:32
    picture, that the ICP is doing bad
  663. 00:27:34
    things to the most precious real estate
  664. 00:27:37
    of the brand.
  665. 00:27:40
    So let's say we have about, you know,
  666. 00:27:42
    we're almost out of time, but to wrap
  667. 00:27:45
    up, do you, in your, how do you,
  668. 00:27:48
    maybe Dr. Rajaj, you can tell us a
  669. 00:27:49
    little bit, it would give you a case,
  670. 00:27:50
    a cute liver fillet patient comes in,
  671. 00:27:52
    walk us through a little bit about how
  672. 00:27:55
    you want to manage that patient. Are
  673. 00:27:57
    you still sticking to non-invasive ICP,
  674. 00:28:00
    or are you doing invasive ICP? Sure,
  675. 00:28:02
    yeah, no, I think we're, so, you
  676. 00:28:04
    know, just with a hypothetical liver
  677. 00:28:06
    failure patient. I think our approach
  678. 00:28:08
    is predominantly non-invasive, we use
  679. 00:28:10
    pure plumbetry for farm to
  680. 00:28:13
    be two hours, Typically by bedside
  681. 00:28:15
    nursing. and then transcranial doctor
  682. 00:28:17
    assessments of, you know, estimated
  683. 00:28:19
    ICP, aboard every day, once or twice a
  684. 00:28:22
    day, if these are normal, that's
  685. 00:28:24
    reassuring. And then, you know, we
  686. 00:28:26
    just continue to observe the patient.
  687. 00:28:28
    If the pupulometry in particular
  688. 00:28:30
    suggests a sluggish and non-reactive
  689. 00:28:32
    pupal, that sets up alarm bells. We
  690. 00:28:34
    typically get an urgent CT and, you
  691. 00:28:37
    know, more often than not initiate
  692. 00:28:39
    osmotherapy or other measures to control
  693. 00:28:41
    ICP When your CTDR pupulometry is mildly
  694. 00:28:46
    abnormal, we just continue to observe
  695. 00:28:47
    the patient. I, you know, in the
  696. 00:28:50
    specific sitting a liver failure, I
  697. 00:28:51
    think what is most important is, you
  698. 00:28:54
    know, make a decision about transplant
  699. 00:28:56
    quickly. And then if you made a
  700. 00:28:58
    decision that that person should be
  701. 00:28:59
    transplanted, get on it right away.
  702. 00:29:01
    You know, I think sitting on the
  703. 00:29:02
    patient is, you know, what sets up the
  704. 00:29:05
    patient for something bad to happen. So
  705. 00:29:08
    I think just speeding up the process,
  706. 00:29:11
    If you've decided to go ahead and
  707. 00:29:13
    transplant, it's probably the most
  708. 00:29:15
    important thing.
  709. 00:29:17
    Great, so that's it for today. Thank
  710. 00:29:20
    you so much again, Dr. Rajuji, for
  711. 00:29:24
    participating in this NCS podcast,
  712. 00:29:27
    Master Class. Thank you so much,
  713. 00:29:30
    Krishna. You're welcome back anytime.
  714. 00:29:32
    Thank you so much, John. Thanks,
  715. 00:29:33
    Stefan.