Intranasal midazolam hit the scene with a lot of hype. It was fast. It was easy. As a result, it was quickly adopted into many pediatric seizure algorithms. However, the more I think about it, the more I realize that intranasal midazolam makes absolutely no sense in the management of status epilepticus.
I will come back to that thought in a moment. First, to get a sense of the evidence, let’s start with a paper. When managing status epilepticus, you should obviously use an IV if you have one, but what should you do if an IV is unavailable? This paper is a systematic review and network meta-analysis of non-venous medications for convulsive seizures.
Arya R, Kothari H, Zhang Z, Han B, Horn PS, Glauser TA. Efficacy of nonvenous medications for acute convulsive seizures Neurology. 2015; 85(21):1859-1868. PMID: 26511448
This is a systematic review of various non-venous routes and drugs used in RCTs for the treatment of convulsive status epilepticus. They performed a network meta-analysis as a way of comparing the various treatments indirectly, because many had not been compared head to head.
They found 20 RCTs of non-venous anti-epileptic therapy. Patients ranged from newborns to 102 years old. (In other words, there is a lot of heterogeneity in this data, and we need to be very cautious in our conclusions). The overall methodologic quality was low, with 16 of the 20 studies being unblinded.
Unfortunately, many of the studies did not overlap, so many of the routes and drugs could not be directly compared to each other.
For the outcome of seizure cessation in less than 10 minutes, intranasal and buccal midazolam are better than either rectal diazepam or sublingual lorazepam.
For the outcomes of seizure cessation in less than 5 minutes, intramuscular and buccal midazolam look about the same.
In terms of time from drug administration to seizure cessation, intramuscular midazolam was the fastest (2.1 minutes), followed by intranasal midazolam (2.4 minutes), and then rectal diazepam and buccal midazolam (both coming in around 4 minutes).
In terms of time from arrival at the hospital to seizure cessation, intramuscular midazolam was again the best (3.8 minutes), and was superior to both IV diazepam and IN midazolam by about 2 minutes. This is partially because intramuscular midazolam is the fastest to draw up and give (0.8 minutes as compared to 1.2 minutes when used intranasally).
The underlying data here is far from perfect. In general, I think your choice of benzodiazepine matters a lot less than using the appropriate dose and getting it in quickly.
However, reading this study made me question my use of intranasal midazolam for seizures. The evidence for intramuscular midazolam is somewhat more robust. (Silbergleit 2011) I think intramsucular midazolam looks a little better than intranasal midazolam when compared head to head in this study (although the data is too weak to be sure) and the intramuscular route is probably more reliable than the intranasal route in a critically ill, seizing patient.
Intranasal medications are great. I use them all the time for analgesia and sedation, but their value lies in their ability to provide medications rapidly without the pain of a needle. Limiting pain is an important goal in emergency medicine, whether we are talking adults or kids, but when it comes to treating seizures, I think we might have lost sight of the forest for the trees. These patients are unconscious. They can’t feel pain. There is absolutely no reason to choose the potentially less reliable intranasal route over the tried and true intramuscular route in the management of seizures.
Honestly, considering that the patient can’t feel pain, it might even be better to just start an IO and give the midazolam that way, rather than messing around with intramuscular dose, but I haven’t got quite that far myself. (Harms from an IO will be higher than with IM, but it does give you vascular access which will be necessary to give other anti-epileptics.) In order to convince me that intra-nasal midazolam is a good idea, I would at least want to see a large non-inferiority trial comparing it to IM midazolam, but considering how good and how safe IM midazolam is, I am not sure such a study is worth the resources.
Intramuscular midazolam is probably supported by better evidence than intranasal, is likely more reliable, and considering the patient can’t feel the injection, should be our first line agent if we don’t have an IV in status epilepticus.
Arya R, Kothari H, Zhang Z, Han B, Horn PS, Glauser TA. Efficacy of nonvenous medications for acute convulsive seizures Neurology. 2015; 85(21):1859-1868.
Silbergleit R, Lowenstein D, Durkalski V, Conwit R; Neurological Emergency Treatment Trials (NETT) Investigators. RAMPART (Rapid Anticonvulsant Medication Prior to Arrival Trial): a double-blind randomized clinical trial of the efficacy of intramuscular midazolam versus intravenous lorazepam in the prehospital treatment of status epilepticus by paramedics. Epilepsia. 2011;52 Suppl 8(Suppl 8):45–47. doi:10.1111/j.1528-1167.2011.03235.x
Morgenstern, J. Is there any reason to use intranasal midazolam for seizures?, First10EM, May 18, 2020. Available at: