1. What are the essential investigations for a patient presenting with refractory status epilepticus?
2. Once traditional treatments (propofol, benzodiazepine infusion, and barbiturates) have been exhausted what should be our next line therapies?
2. Where should these patient’s be managed and what is the minimum monitoring standard?
This month’s paper is a fantastic article from the NEJM from this year, pertinent to all Intensivists – https://www.nejm.org/doi/full/10.1056/NEJMra1708712
Here is an article I read during our two recent patients being on the ICU at the LGI.
Its an easy to read commentary on less common therapies and drug regimes such as pyridoxine, ketamine, ketogenic diet, ECT, immunotherapy and deep brain stimulation etc.
In terms of specific investigations, I think from our recent experiences of patients with refractory/super-refractory status epilepticus (RSE) my threshold for an urgent MRI (in the absence of abnormal CT findings, and in those without a history of epilepsy) has certainly dropped, and early consultation with neurology for specific CSF antibodies from the initial, rather than repeat LP are two aspects I am more conscious of. I would hope that the diagnostic superiority of MRI may allow us to try more specific treatment therapies such as immunotherapies, but also allow us to better prognosticate as the morbidity and mortality associated with SE is largely dictated by the underlying pathology rather than duration of seizures. This was nicely demonstrated by the patient B mentioned (known epileptic rather than new-onset refractory status epilepticus (NORSE)) who walked onto the ward a couple of weeks ago, completely fine on their way home despite at least 72hours in RSE and 2 week ICU stay.
Our treatment strategies (in addition to therapies directed to the postulated cause) have traditionally consisted of an escalation from propofol infusions to the addition of midazalom (although the utility of benzodiazepines reduces significantly as seizure duration continues) and then thiopentone (in addition to augmentation with phenytoin, Keppra +/- Valproate), with the inherent risks that increasing doses of these infusions carry (PRIS, dyskalaemias, CVS instability to name a few) . Therefore it seems sensible to have both a targeted, demonstrable endpoint (in LGI we have the advantage of being able to demonstrate burst suppression on our CSA and 2 raw EEG traces, although I am uncertain how useful they are to us non-neurologists for anything inbetween), and consideration of alternative therapies. The problem with the latter is that because of the rarity of RSE the evidence base at this point is weak, and things we have used in true “hit and hope” fashion include magnesium infusions, methylpred, locosamide & topiramate. I have never personally seen or used Ketamine in this cohort of patients, but it does seem to have undergone a reinvention as a brain friendly drug (at least since my primary FRCA) and crop of fairly regularly in the recent literature. I must admit our familiarity with it coupled with the administration by infusion does make it seem like the more attractive of the lower evidenced options. As always, interested to read other people’s thoughts.
Sorry for the delayed reply but it has taken me a little while to get the chance to read the articles in depth.
I was the clinician looking after one of these patients and had to resort to “phoning a friend” (thanks Justin and Paul) to help navigate the various non-evidence based therapies as all the usual stuff had failed.
I agree with all the points you made James and I would add another few having reflected upon the cases and the articles presented.
1. In cases that don’t fit with infectious cause (and probably even in those that do) I am going to start collecting an extra bottle or two of CSF. The number of times we have had to try and get a second sample or eek out the samples we already have in order to perform the immunological tests has increased.
2. I will start adding TFTs to my initial battery having read the articles
3. I will have a lower threshold for doing the investigative fullbody CT once the commonest causes have been ruled out.
4. Most importantly, I think I will use ketamine a lot earlier. I too had a hang over concern about ketamine in “heads” following the FRCA but the limited evidence and the suggested pathophysiological benefits (GABA receptors have been internalised making benzos less effective, thus NMDA action preferable) have persuaded me I should use this earlier.
I agree about using ketamine early, and like Martin says I can’t see any reason not to give magnesium by infusion, with deep tendon reflexes used to ensure safety.
I found the same article that A circulated, and one thing I think is important is that we should be regarding our anaesthetic agents by infusion as being the GABBA agonists and then selecting non-GABAergic anticonvulsants on top. The most potent GABA agonists are our drugs – benzo, barbiturates – so employing lamotrigine, gabapentin, valproate makes less sense as we will have likely already saturated the GABA receptors. Also, if the functional GABA receptors go down all GABAergic drugs will become less potent, so using Keppra, phenytoin, pyridoxine, topiramate with non-GABergic effects should probably be the drugs to use, in addition to drugs targeting glutamate including ketamine and magnesium.
One issue I don’t know the answer to, is when to say enough is enough of one agent. Whilst with phenytoin we can measure a therapeutic level, and there are agreed ranges, this isn’t the case for most agents including midazolam, STP etc. When have we given enough and need to resort to another additional agent? Whilst the elegance of EEG monitoring can’t be denied, in both the cases under discussion I was never sure that burst suppression was ever convincingly achieved with STP (or not for very long). The maximum doses in the literature for STP gets quoted as up to 15mg/kg/hr, but do we dare to go this high, I don’t know.
Investigate the cause, and this includes LP, MRI, whole body CT, and where there isn’t one, consider immunosuppression. Also, in experimental epilepsy hypothermia exhibits anticonvulsant effects and in theory should be neuroprotective. In reality this is likely to need either endovascular cooling or large area surface cooling with a water circulating system.
High dose barbiturates, hypothermia, what could go wrong!
I do recall using Ketamine, it did appear to work.
Some of you have emailed me directly to ask for specific doses and so here is a review published less than a month ago on this topic (it was fortuitously emailed out by Critical Care Reviews a few weeks ago) that has collated doses from the most recent observational studies and case reports. I don’t expect there will be any higher level evidence anytime soon, as several recent planned RCTs in this domain have been stopped due to failure/inability to recruit because it is an unusual presentation.
E‘s question of when is enough “enough” is clearly very difficult but I do worry with reported but high ranges (such as 15mg/kg of STP) that we introduce significant risks to treating a symptom (i.e RSE) with ill-defined but probable harm, of a pathology and not necessarily the disease itself.
I think you have made some logical comments about choice of agents. You will remember back to when we consulted Morgan Feely on this topic for our book. He was clear that in the vast majority of cases of refractory SE, the reason that seizures had not been controlled was because inadequate doses of typical anticonvulsants such as phenytoin had been used. I would certainly wish to avoid rapid escalation to polypharmacy with infusions of ketamine and magnesium etc before we have got the basics right. I agree that prompt investigation of possible underlying pathology and consideration of immunosuppression is important.
The maximum doses of STP suggested by the literature seem ludicrous to me and I won’t be an early adopter!
On monitoring levels…
Is it common practise to correct phenytoin levels in cases of hypoalbuminemia or uraemia?
I sought some advice from one of the neuro critical care pharmacists about this and its scary how high the corrected level can be (in one case I remember it was over 40mg/l when corrected).
The effect of hypoalbuminaemia can be significant, particularly in patients with an albumin <20
Phenytoin is >90% protein bound in the average patient – only the remaining 10% of free drug is pharmacologically active.
Low albumin states will increase the proportion of free phenytoin.
The laboratory measures total phenytoin levels rather than free phenytoin – the therapeutic range is based on the average patient.
In a hypoalbuminaemic patient the level should be adjusted so it can be interpreted more effectively – you use a similar calculation to the one the labs use for adjusted calcium.