Critical Care Medicine - List
http://www.pitt.edu/~crippen/

Intracranial Pressure Increase and what to do about it

Presented by Louis Brusco


Louis Brusck:

I had a 27 year old gunshot to the head, adm 12/23 with GCS of 3 on presentation, increasing to 7 with resuscitation. CT showed bullet & bone fragments intracerebrally, occipital entry & ext. Inubated, stabilized, taken to OR 12/24 for debridement & closure of wound (6 hrs after presentation). After OR, repeat head CT now with loss of ventricles, no shift. In SICU given increased doses of mannitol, hyperventilated, also given steroids (per neurosurgeons), and furosemide. Serum osmolarity to 320. Repeat head CT 12/26 with worsening edema with total efactment of sulci. Depite this, patient arousable after wakeup from propofol and follwoing commands, opening eyes.

Over the course of yesterday, the patient had cardiorespiratory decompensation. Unable to evaluate neurologically because of mandated switch from propfol to lorazepam after 48 hr. Temp. to 107 tothis morning, bloodpressure dropped, and patient had v-tack/fib arrest from which he could not be resuscitated. Pre arrest, his pupils had grown large and much more sluggishly reactive.

I feel that this patient herniated, and that, in retrospect, the only thing that we could have done for him was place an ICP monitor and place him in a barbiturate coma when he plateaued. My neurosurgeons steadfastly refuse to place ICP monitoring lines, saying that the literature doesn't prove that patients do better with them, and that they get infected. They also use hyperventilation for days on end despite indications that it only works for 12-24 hours and give steroids for cerebral edema, and never give enough mannitol (this time, I raised their 12.5 mg every 8 hours to 50 gm every 4 hours). These are otherwise excellent neurosurgeons who I trust with my family;s lives, but we don't see eye to eye on increased ICP.

What do you do at your hospital? Can you point me in the direction of hard-to-refute clinical evidence that ICP monitoring is beneficial? I hope to argue the point yet again with them at next Wednesday's M&M. Thanks.

Steve Curry:

Admittedly, with rare exception, bullets that cross the midline of the brain are not compatible with nonvegetative survival, despite initial encouraging signs and agressive therapy. Therefore, there will probably never be a study demonstrating that more aggressive controlling of IC hypertension will improve outcome.

Tom Bleck:

Unfortunately, the study showing that ICP monitoring leads to interventions that improve outcome was never done, and never can be done (those of us who might be in a position to do it would not consider it ethical to withhold it; at least I wouldn't). The notion that ICP should be controlled to a level of about 20 mmHg comes from studies (e.g., those by Don Becker) that showed survival to be better in patients whose ICP could be kept below that level; this does not answer the question of whether these people were more likely to survive anyway. To really study this question would require monitoring patients but keeping the data from their physicians on a randomized basis, unless one wanted a study the size of GISSI.

We use the Camino intraparenchymal monitoring system, which employs a fiberoptic transducer instead of a fluid filled column and has a vanishingly small risk of infection. We use cefazolin prophylaxis while the catheter is in place; this was a move from vancomycin because the hospital was having trouble with VRE. We do have occasional ventriculostomy infections, but since we don't use them in trauma patients, this has not been an issue. The down side of this monitoring technique is that one cannot withdraw fluid to lower the ICP, but as I'll relate below this is not a serious problem for us.

The steroid issue was resolved in the 1970s be Larry Marshall, who demonstrated in a RCT that steroids worsen outcome in head trauma. His chapter with Randy Chestnut in Allan Ropper's _Neurological and Neurosurgical Intensive Care_ (3rd ed, 1993, Raven Press), is an excellent summary.

Our policy is to use mannitol and hyperventilation only as short-term temporizing measures. The hyperventilation question has been well studied over in Richmond; I refer the inclined to read the papers by Muizelaar et al on this point. The mechanism of action of mannitol is currently unknown, by the way, and the utility of osmolality measurements for its dosing is unproven. Another example of the Firesign Theater effect (from their album, Everything You Know is Wrong).

When simple measures (sitting up, keeping the head in neutral position to improve jugular venous return, keeping the intra- thoracic pressure as low as practical (but _not_ volume restric- tion), sedation, and pain relief) don't work, we generally pop the top. Craniectomy, IMHO, is much better tolerated by the patient that high-dose barbiturates.

Louis Brusco:

In this instance, it was a single GSW to the left occiput that had an exit wound only 2-3 cm lateral to the entrance wound. It is unknown whether the bullet traversed the cerebrum then exited, or whether it was confined to a single side. What would you do in this instance? Is there anyone who NEVER uses ICP monitoring?

Kelly Randolph RN:

From a nurse's perspective I've always hated ICP monitoring because it then gives me another peice of equipment to watch. Then there's the tendency to "treat the numbers" rather than the pt and his neuro signs. Its been my experience that the tx is the same whether invasive monitoring is done or not. With a monitor in place its more suspenseful and frantic the pressure goes up, up, up........ We use steroids here most of the time, and barby coma's too. You bump your head, you're dead.................

David Crippen:

Mannitol, can rapidly decrease intravascular volume so rapidly brain shrinks and pressure decreases such that any tampanoded vessels have a renewed incentive to start bleeding. IMHO, mannitol probably reaches the point of diminishing returns quickly. I only use it for emergency rescues when patient is destined for a quick trip to the OR and I am trying to buy a little time.

Unless injury is completely homogeneous, hyperventilation constricts vessels where autoregulation is still intact, does nothing to vessels in the damaged area, good blood shifts to bad brain. Steroids Don't decrease edema unless it is of inflamatory origin. May have adverse metabolic side effects.

As far as I know, there is no evidence that Barbiturate Coma improves outcome from brain coma when instituted after the injury. If you could predict the injury, placing them in Barb Coma probably helps.

Switching from propofol to lorazepam is a VERY bad move. Lorazepam is a LONG ACTING AND NON-TITRATABLE CNS DEPRESSANT. Some say it never goes away.....it just keeps on truckin' like that little bunny.

Temp. to 107. Hmmmmmm.......compression or anoxic damage to the thalmus...this is definitely not good. I would not sell this patient any long playing records.

V-Tachy/Fib arrest from which he could not be resuscitated........I think this was a predictible end, and the probability was high regardless of your administrations. Again, IMHO, I think the only thing an ICU monitor would have given you is a number you couldn't really do anything about. However, he was not at Ben Taub......did you consider transfering him? :-) :-) :-) :-) :-)

Mike Rie:

Dear Louis: In Kentucky all head injury in Emergency room with GCS<7 not due to alccohol drugs or iatrogenic NMBs gets a ventricular drain, not a surface monit or. The bullet type case as you described gets it in the OR. Sometimes we unroo f the skull for a while. Our primary strategies are drain the brain as we maxim ize the cerebral perfusion not just CPP. The local neurosurgical strategy is ne ither heavy on hyperventilation or dessication. Results are impressive overall. We are real fast on the intraventricular catheters because when you daly, brain swells fast and slit ventricles are nearlly impossibe to cannulate. We go to barbs when drainage fails.

Gordon Doig:

Here in Ontario there are approximately 7 Level 1 trauma centres. Some (approx half) so infrequently use ICP monitoring that it might as well be called 'NEVER'. We just finished an epidemiological study looking at outcomes of patients with ICP monitoring. We controlled for ER type stuff such as TRISS (as opposed to APACHE), MAIS head, GCS (modified for intubation etc) using a 5,000 patient trauma (ISS12) database.

We expected to have findings similar to Gore's when he looked at the use of the Swan (more severely ill patients require Swans so thus patients with Swans have a higher severity-adjusted mortality rate). Our findings were however VERY interesting. Our control patients were very conservative (definately less ill than the ICP patients) and we still found a survival benefit of ICP monitoring! We are not sure what 'interventions' in reaction to increased ICP are being used, but this finding can definately be used to support the rational for an RCT (in Ontario at least, since half our hospitals dont use ICP monitoring routinely).

We hope to publish our findings in the Journal Trauma soonish. To us, this is the strongest argument for the use of ICP monitoring, but we still conclude that an RCT is needed to provide 'Sacred Calf' level of evidence.

Malcolm Fisher:

We dont tend to in gunshot wounds as most (?al)l are self inflicted and i believe a pretty definite expression of autonomy. I have a similar approach to self immolation.( Maybe Gun Laws would be more effective?) I think Thomas Bleck's summary of what is good and what can be proven is excellent. Like most critical care it is the package that saves lives and picking out one bit to study often fruitless. I like to interpret the Becker/Miller stuff as showing the efficacy of the package.

The two questions that fascinate me are why we dont use hypertonic saline instead of mannitol as it shifts more water than mannitol(excellent account in CCM State of the Art by Prough a few years ago) and the role of NSAIDS. There was a poster in Brussels a few years ago that showed unimpaired outcomes in patients with raised ICP refractory to barb coma. This is usually a death sentence in our place. The literature search I did subsequently suggested they have a place but our Neurosurgeons are worried about bleeding and in the one hopeless case I tried the renal hit was even bigger than the NSAIDs renal hit in most critical people.

Lou Brusco:

The issue of expensive sedatives is one area that I have been unable to change pharmacies feelings. This guy was getting 80cc/hr of propofol for 48 hr, and they refused to give me any more. This is depsite morphine 10 mg/hr (on which he was following commands on wake-up testing).

I may be wrong, but I feel that this guy would have herniated a lot faster if I hadn't given him 100 gm mannitol every 4 hours with lasix. One could argue that he would have died anyway, but I can't see trying to perfuse a brick when you could perfuse clay (bleeding considerations aside).

My experience is colored by one patient who was in an MVA and had bad closed head injury with diffuse edema, with also a bad pelvic frature that would not stop bleeding despite embolization. WHen taken to the OR, adn ICP monitor was placed (not at this hospital, obviously). During the case, she plateuaed to 50 mmHg and was placed in a barb coma. She was kept in it for 5 days until we could wean it off and keep her ICP < 20. She was in a vegetative state for a few months, and I lost track of her. 9 mos later a girl came to the SICU and introduced herself. Except for a slight limp, she seemed intact. It was that patient.

George Skowronski:

This business about measuring various things not being shown to improve outcome - this worries me. A measurement can't affect outcome - it's only how you respond to the measurement that can possibly make a difference. Nobody has ever shown that measuring blood pressure with a sphygmomanometer improves outcome, but there's a pretty big body of evidence that treating hypertension does. I think we're asking too much of evidence based medicine to show that a measurement of itself improves outcome. Any views?

Stephen Streat:

After OR, repeat head CT now with loss of ventricles, no shift is a bad sign but not unexpected. We would do SEPs (they would have been normal if your patient was able to open eyes and obey command) and decide if we were prepared to go to thiopentone coma. If so we would insert fibreoptic brain tissue monitor (no sepsis, minimal bleeding risk in patient without coagulopathy, good device but no CSF drainage possible - doubt you could have entered the lateral ventricle in your patient safely anyway). Our neurosurgeons are more cooperative (=`tame(d)") than yours. It helps that technically these are our patients not theirs in terms of strict ownership. We use this technicality to our (and the patients) advantage at every turn. If we ask `please insert Camino' they insert Camino. Elsewehere in NZ where there are no neurosurgeons the ICP monitors are inserted by intensivists (eg Jack Havill in Waikato). They are (apparently) no more tricky than PA caths or perc trache and we are happy to do those.

We do not use mannitol except for pre-op rescue. We use 4Msaline - it makes water handling easier to manage- they do not get a big osmotic diuresis and hypovolaemia, followed by volume loading, follwed by increased ICP, followed by mannitol - if you know what I mean. would still aim for serum osmolality around 300 initially, up to 320 with ICP monitor in place.

hyperventilation is a two edged sword, lower ICP, worse cerebral oxygenation, we have compromised at paCO2 30-35 mmHg and don't (yet) monitor jugular venous bulb oximetry.

The repeat head CT 12/26 showed worsening edema with total efacement of sulci. Depite this, patient arousable after wakeup from propofol and following commands, opening eyes. This is interesting but why stop anesthesia when CT shows severe oedema ? We would monitor CNS by neurophysiology and keep patient curarised and sedated - probably ICP monitor would have shown ICP 25 or so and we may well have had him on thiopentone infusion (100mg/kg load, 2 mg/kg/hr infusion - aim for burst suppression EEG). On the supply side we keep CPP above 70 with noradrenaline and colloid (= thicker runny stuff with oncotic properties). Reducing core temp to about 35 is quite inocuous for days (lower temps associated with WBC margination, leucopaenia and sepsis) and will usually give very favourable mprovement in cerebral compliance.

Over the course of yesterday, the patient had cardiorespiratory decompensation.

If your neurosurgeons steadfastly refuse to place ICP monitoring lines, how about placing them yourself ? We find that direct action works pretty well here - it helps to have cooperation but if not - hell, what have you and the patient got to lose ? No RCTs it is true but there are no RCTs that show that keeping peoples PaO2 over 70 is better than over 50 either. Brains certainly work better when their perfusion is maintained (GCS rose from 3 to 7 with resuscitation as you said) and ICP monitor enables fine control of cerebral perfusion pressure at least.

We treat ventilate 120-130 patients with AIS4 or 5 head injury per year. All get CT scan, paCO2 30-35 (initially), extracranial fix (including colloid-biased circulatory resuscitation and liberal inotropes to keep MAP 90+ as default mode). If CT shows no mass effect we examine clinically after extracranial fix. Other patients get SEPs wiuthin 24 hours. ICP monitoring is reserved for patients with favourable SEPs and CT or operative findings of oedema. Osmolality is kept 290-300 in patients without ICP monitors, may be higher in monitored patients. Thiopentone and hypothermia (and THAM, hyperoncotic albumin etc) are reserved for patients with ICP20 despite osmolar control and curarisation and generous sedation/opioid. Old fashioned huh.

In the words of my old Scottish boss - Matt Spence - the founder of intensive care in NZ (1958) `yee just gottee tell'em.' Sometimes these issues are not resolvable by reasoned argument - it boils down to who is sitting on the mushroom and who is standing below asking `how can you make words mean so many different things '.

Some refs from Stephen Streat:

Tom Bleck:

Mike Rosner at UAB has done some work on induced hypertension for head trauma. The basic arguments are that (1) CPP is important, perhaps more important that the absolute ICP in patients with global processes (as opposed to focal ones which will shift the diencephalon), so one can maintain CPP by raising MAP (I apologize for the abbreviations in advance); (2) raising MAP will shift the autoregulatory curve to the right, and by producing vasoconstriction will lower cerebral blood volume and therefore decrease ICP; and (3) one potential mechanism of mannitol's ICP effect is to increase preload, which might, possibly, conceiveably, potentially raise MAP and thus act like (2) above. Given these arguments, Mike opts to skip the mannitol and just use norepinephrine to maintain the CPP. This is the subject of the J Trauma paper someone provided a reference for earlier (Apologies to the provider for my lapse about your identity).

The real value of this work may be the reminder that CPP should be maintained. Whether induced hypertension is the best way to do this remians, in my mind, to be demonstrated.

In a previous life, before I bought Camino monitors (and ran them out of the EEG lab, since that was my only option), we sometimes used EEG burst-suppression as the marker of adequate barbiturate dose. This goes back to some 20+ year old studies by Neal Kassell, showing that the maximal decrease in ICP and cerebral oxygen consumption occurred somewhere around the burst suppression threshold. This makes physiologic sense as well, since this EEG finding is the threshold for complete inhibition of synaptic transmission. In burst-suppression (e.g., about 10 sec of 'flat' EEG followed by a brief, bisynchronous burst of relatively high amplitude activity), a little more barb will make the record flat. One doesn't know how deep the water is after falling off that edge, but there is probably no ICP or CMRO2 advantage to getting completely flat.

This is not quite as simple as it sounds. One night, I instructed the house staff on how to recognize and maintain this pattern in a teenager with herpes simplex encephalitis (doing it for control of status epilepticus, which is not a reasonable indication for a burst-suppression EEG, but we didn't know it at the time; more on this another day). The next morning, I came in to check on the boy, and was told that despite many grams of pentobarbital they were unable to get the bursts more than 6 seconds apart. A quick review of the tracing showed that the EEG had been 'flat' for several hours, but that they had been trying to eliminate the ventilator-induced movement artifact. The EEG did not recover any activity for several days, but eventually both it and the patient recovered.

Richard Piper:

Further to my last message, we in Adelaide were also using inotropes for head injuries to increase CPP (usually adrenaline). Our experience was that ICP tended to rise, but CPP rose more. Interestingly, SjO2 also moves in the right direction when this happens, and seemed to us to vindicate the approach. John Myburgh at the Royal Adelaide Hospital wrote a nice clinical protocol for the registrars which included the use of inotropes, mannitol and how to play the SjO2 into all this. He's presenting some of the Adelaide stuff at the Brussells meeting, for those who are going.

George A Skowronski:

In the context of head injury, is an increase in SjO2 _always_ a good sign. For instance if your therapeutic intervention caused an increase in the heterogeneity of flow and hence mismatch between the distribution of oxygen demand and supply in the brain (ie shunt) might it not increase SjO2 without increasing total brain oxygen extraction. Presumeably a similar mechanism causes the increased mixed venous O2 seen in sepsis.

Avi Roy-Shapira:

Admittedly, with rare exception, bullets that cross the midline of the brain are not compatible with nonvegetative survival, despite initial encouraging signs and agressive therapy. I beg to differ. The exceptions are not so rare. I once saw a women who was shot in the left temple. She walked into the ER, and was yelling at the cops who brought her in. The bullet was lodged in the right temple, and CT clearly showed the track through the brain. She did fine, and never lost a beat.

Admittedly, this is a rare case, but I have seen several patients that had transcranial bullet wounds and are now alive with minimal neurological deficit. The deterioration to chronic vegetative state is, in my experience, less common. Most of these either die or recover (most, of course, die) The question is really what data is there to support ICP monitoring (and its more aggressive cousins such as Jugular venous sampling) in ANY head trauma. I know of no RCT's in this regard, however, it does make management easier. More data ought to account for something, and easier management is in itself a goal of intervention. We all use SG catheters, and there too the data is lacking. I believe a good intensivist will be able to get by without a SG with little difference in mortality.

Finally, the 48 limit on propofol in this setting is crazy. This is exactly the type of patient where propfol should be used. Had propofol been used, deteriorating neurological status could have alerted the team to the worsening ICP. The point is, in my view, not whether ICP monitoring changes anything, but that one has to have SOME way of monitoring neurostatus. If a long acting sedative is used, then ICP monitoring is a must. If one can follow the patients by physical exam, ICP monitoring is much less useful.

(Re- lorazepam)....This is outrageous.An alternative, which I use rather commonly, when I get a negative reply, is to ask the pharmacist (or whoever made the policy) to sign a note in the chart saying that the drug, procedure, apparatus etc, is not indicated in THIS case.

I have found the latter approach very useful in getting recalciterant radiologists to do pulmonary angiograms at 2 am, or aortograms on trauma victims with widened mediastinum. I tell them that I see them as my consultants, and that I fully trust their judgement on the matter, I say that I must insist that they write a note in the chart saying that there is no indication for the procedure, since I must protect myself. Invariably, they agree to do the procedure.

Beaurocrats loath taking personal responsiblity. The very nature of beaurocracy is that one acts by rules, regulations, and orders from above. Nobody is reponsible. Force them to take personal reponsibility, and they will yield.

There is yet another approach. Instead of having a cap on the use of a drug, nominate someone as an authority and have him or her sanction the use on a case by case basis. We use this approach for most advanced antibiotics, and also for Albumin. You need to authorize more than one person, so that the authorization will be available all the time when needed.

In my hospital Albumin was used rather indiscreetly. I was one of the three appointed guardians. Since we were nominated, Albumin use had dropped 80%, yet all who needed it, on bona fide indications, were never denied. Using this approach, one can save money, without getting into the rediculous situation Lou has found himself in. If anyone is interested, I can provide details of the mechanics.

Lou Brusco:

I was, in fact, one of the appointed "guardians" for propofol. However, in anticipation of the Gingrich-Dol medicare cuts, which will have a disproportionate effect on NYC hospitals, on top of state Medicaid cuts, we are looking at trying to cut $30 MILLION from our budget next year. This was one of the more resaonable cuts, but, in these times, no exceptions are allowed. In this case, with the guy getting over $700/day worht of propofol, I was forced to go with the $30/day lorazepam. It is going to get worse.

George Skowronski:

In the context of head injury, is an increase in SjO2 _always_ a good sign. Absolutely - the SjO2 has to be interpreted in exactly the same way as MVO2, in the clinical context of what's going on. A rising SjO2 in the context of worsening CPP or whatever could be sinister. Even more importantly, brain dead patients have terrific SjO2's due to extracranial contamination, etc. In practice,though, I found it not too hard to tie things together looking at the overall clinical picture.

Mike Rie:

A New years reflection on ICP management comes to mind after reading the though tful posts of Thomas Bleck and Crippen's need for an excuse to checkin on the Safar work on induced hypertension for ICP. We have the 8th most active head injury trauma service in USA here and lots of experience as previosly stated with draining brains, barb coma, etc. My reflections are threefold.

  1. We never use propofol and drainage gets us through in a cost effective manner for a large number of case. We cover caths with nafcillin at striclly change catheters and closed drainage systems btween 5-7 days. Nafcillin coverage is used but I have no idea about its efficacy. I think we do a a lot of CT scansand I have seen some of therse patients get as much as 20 follow up scans which probably represents more money than days of propofol but I have not done the arithmetic.

  2. There is nothing more expensive, morbidity producing and probably of miniscule benefit than barb coma for elevated ICP. To boot when you decide it fails, you invariably have to do cerebral flow scanning to document brain death or wait for several days till all the pentobarbital washout. The neurosurgeons manage this therapy, always require PA line, inotropic support and occaisionally experience multiorgan failure with total body blue mottling, lactic acidosis and low cardiac output. These resuscitations with volume repletion, calciumepineprine etc invariably fail. Twice now I have seen but not reported in the literatue that institution of small meterage coil conventional dialysis incited terminal fatal surge of the ICP before any fluid had been removed and occurred within 5-10 minutes of beginning the dialysis.

  3. We like Bleck have tried craniectomy when the rest of the therapies was failing. I am now thinking that we should try a small number of patients earlier and see what comes of it. Avoiding barbs will avoid need for PA lines, followup daily head CTs and decrease excess ICU stay from complications of barbs. I also neglected to mention that in our hands barb coma patients suffer a greaterincidence of infectious complications. We now feed almost all head injury patients by small bowel enteral nutrition with regular metabolic assesments. Thesepatients have the highest stress factors we see and feeds start at day one so nutritional depletion does not appear to be the principal culprit for excess infectious complications with the barbs. The stress repletion factor decreases slightly with barbs and we adjust the feeding accordingly.

Avi Roy-Shapira:

I find it hard to believe that one patient getting propofol is going to make a dent in your 30 million budget cut. To cut 30 millions effectively without affecting patient care, you need to get rid of middle management. This is where the big money is saved. The great majority of patients do nicely without propofol. We use it VERY rarely. We would however in this patient, for as long as necessary. IMHO, what one should guard against is not the use of an expensive drug for the rare patient for which it is needed, but against the indiscriminate use in patients who do nicely without it.

I don't envy your situation, you were left without any way to monitor the patient, and you lost him. He would probably have died anyway, but you don't know that for sure. Your neurosurgeons are not without blame, of course. As you say, it is getting worse. It seems to me that you Americans have finally found a way to take the best medicine in the world, and ruin it. I read a theory somewhere that big successful systems always commit suicide in the end. Witness the Roman Empire. So perhaps this is the case for American Medicine. Having been trained in the US (I hold ABS certification) in better times (the eighties) I feel very badly about this.

David Crippen:

I think Lou asked earlier how I personally dealt with head injury.

We do not use ICP monitoring. We will put in an intra ventricular drain if the vents are big and there is a reasonable likelihood of hitting one with a catheter.

I sit them up in bed, tape their forehead to sides of the bed in a face forward position to insure good jugular drainage. I sedate them adequately if needed (not with lorazepam) I promote good diuresis and try to insure that the patient is in negative I/O balance. Our neurosurgeons are fond of a concoction called "Brooks Solution", which is a D5 1/2 normal saline with variable amounts of potassium, lasix and bicarbonate in it. The theory is that is (somehow) buffers the intracranial intravascular space from constitutional fluctuations that might cause either too much or two little fluid in the capillaries. My private opinion is that it is Voodoo and the only practical help is that it does promote diuresis. My public opinion is that it is like Intravenous Chicken Soup......it might help or it might not but it probably doesn't hurt anything and so I am willing to be accommodating to keep peace.

I avoid giving hypotonic IV solutions, keep the serum sodium over 140 or so. Most importantly I keep the blood pressure high enough to get good perfusion through vessels no longer controlled by autoregulation. If the BP gets too low, perfusion pressure decreases and cells are starved for nutrients. If pressure is too high, increased hydrostatic pressure pushes fluid out and increases edema. I was taught that a systolic around 140 was optimum for most. I think there is a mean BP calculation associated with that but I forget the exact number at the moment.

I almost never put people in barbiturate coma for brain swelling. I have never been impressed that there was any practical improvement in outcome. Ditto steroids. I use hyperventilation only for brief periods to get over humps....ditto mannitol. That's about it.

Roger Traill:

I think that you need both to use Barbiturates properly in such cases. The logic behind barbiturates is simply to control the ICP so as to prevent cerebral ischaemia due to brain herniation. The maximal effect is acheived when the EEG is isoelectric but you may not need the maximal effect.

The ICP to aim for is <20torr (perhaps <15 if you have temporal lobe swelling) so we have two end points, firstly the minimum dose of barbiturates to reduce ICP below these limits (regardless of what EEG activity is seen, provied there is some) and secondly the dose when the EEG is isoelectric even if the ICP is NOT below the desired level (you then have to find something else).

It would also seem to me that one of the major objections to their use is the long duration of action however if one uses Methohexitone (which is one of the commonly used barbiturates in ischaemia studies) this would not be such a problem. Despite this I am not aware of anyone who uses it! There is no logic for this. It does not cause seizures in doses that suppress the EEG and is not expensive (come to think of it it might be a good IV sedative in those requiring long term propofol).

As an aside mild hypothermia is very good for head injured patients and hyperthermia is very bad.

Mike Rosner's paper is interesting however if one looks at the literature of Muizellar and Bouma one can see that the benefit depends on whether there is intact autoregulation. If it is then the compensatory vasoconstriction that follows an increased CPP may decrease the ICP however if it is impaired then ICP may increase.

Point 2 is incorrect, the curve is not shifted to the right by raising the MAP rather the patient moves to the right along the existing curve.

Whilst the general view is that a CPP of 70 is the point to aim for induced hypertension should only be done where you have the ability to measure ICP.

Tom Bleck:

This goes back to some 20+ year old studies by Neal Kassell, showing that the maximal decrease in ICP and cerebral oxygen consumption occurred somewhere around the burst suppression threshold. This is covered in one of the references I posted previously - this work was on etomidate but the same message holds.

Stephen Streat:

This makes physiologic sense as well, since this EEG finding is the threshold for complete inhibition of synaptic transmission. In burst-suppression (e.g., about 10 sec of 'flat' EEG followed by a brief, bisynchronous burst of relatively high amplitude activity), a little more barb will make the record flat.

We titrate thiopentone to ICP, EEG (burst suppression max) and blood level (after 100mg/kg/load). We only use what we need to get ICP less than 20 or so. If EEG is Gr IV and ICP is still too high and thio level is high we go to hypothermia (33 degrees ) and THAM. Interestingly we have found that SEPs are resistant to this level of thiopentone - even in one patient we found normal SEPs (central conduction time) with an iso-electric (flat, Prior Grade V) EEG. The evoked potential still gets through even when the spontaneous stuff is flat, flat, flat.

We used to titrate thiopentone to EEG burst suppression in patients with refractory seizures (eg due to high dose steroids, herpes and the like - hey, this was the 70s remember). It was not that difficult - we wired the ICU to the neurophysiology lab and used bedside 8 channel EEG and lab-based CSA. Sometimes it worked - it wasn't that nasty to use either - now we give similar doses for ICP control. Sure, they all need noradrenaline but as long as they are fairly well topped up with volume and don't get intercurrent sepsis the inotrope doses were moderate rather than industrial. (We don't need PA caths to manage this and as for letting neurosurgeons do this kind of therapy - you must be crazy).