For as long as I can remember I’ve been told that it is possible to analyse marrow from an intra-osseous (IO) needle. It’s been taught on lots of APLS courses I’ve attended (Ed – in other words you’ve taught it!), but can it really work? There are many recommendations suggesting good correlation, but those who have read the evidence have suggested that the data in critically unwell patients may not be that robust1,2.
On the very few occasions that I’ve heard of marrow being sent to the lab it has been rejected on the basis that it is potentially too particulate to go through the automated analysers. I would also caution against putting it through your departmental blood gas analyser without clearing that with your lab team first. You can make them very unhappy if you block up the mechanism with bits of bone!
So there are practical reasons why your lab might not accept a sample, but what if they could? Is there evidence that the values obtained would be reliable and thus helpful to your resuscitation practice?
This week we’ve found a systematic review on exactly this topic and the results are bordering on #dogmalysis3. The abstract is below, but as always please read the full paper yourself.
It’s a systematic review of papers looking at the reliability of IO samples as compared to blood samples. This is a reasonable approach to investigating the question, but of course the findings are highly dependent on the quality of the papers found. The authors have performed a comprehensive search across several databases, conference abstracts and by hand searching through references. In my opinion the search strategy is pretty good and I think it’s unlikely that they have missed any key trials.
Which papers did they find?
This is a key question and in this case a surprising one. For a practice recommendation that has been rather pervasive over the last 20 years or so the types of papers found. The authors looked at over 800 papers, but just 27 were relevant to the question. Of those, eight (30%) studies included humans, 18 (67%) included animals and one (4%) study protocol included both. Most importantly to me is the fact that only one study included haemodynamically unstable patients. In other words only one of the papers found focused on the sort of patient I see in the resus room4.
The sample sizes in the studies were small, typically 5-30 subjects. In the human study of critically ill patients there were just 17 subjects.4
It’s also worth noting that the authors of the review have included one of their own studies (that took place in healthy volunteers)5.
How did they compare IO with blood samples.
Each individual paper had its own methods, but in general samples were analysed in the laboratory (though a few were point of care). In terms of agreement it looks like many of the studies used simple correlation to assess agreement, or even just a comparison of means. At St Emlyn’s that sort of analysis worries us as it’s possible to show good correlation despite significant numbers of disagreements which could be clinically important. A few studies used the better method of using Bland-Altman plots which take much better account of outliers6.
The heterogeneity of methods and analysis meant that the authors were unable to pool results as a meta-analysis.
What were the main results?
In brief there is very little evidence that IO samples match IV samples in critically unwell patients. The main finding from this study is really the lack of evidence of anything really. The authors suggest that potassium levels are higher in IO samples, but even that is arguably not a particularly robust recommendation.
Chris Connolly reminded me on twitter that marrow may be suitable for use in blood cultures, and I agree that in my Paeds practice we do this for the septic kids.
The clinical bottom line.
Unless someone out there can find better evidence (I could not) then we should not rely on bone marrow analysis in critically unwell patients. Although the only paper in critically unwell humans suggests that it might have a role for some variables I am unwilling to rely on a study of just 17 patients4.
Having been told (and taught) hundreds of times that we can use IO samples in resuscitation I think this paper is #dogmalysis7, and we love that here at St Emlyn’s. Next month I’m teaching APLS in Virchester and I suspect that I might be struggling in the IO practical session.
- 1.Nickson C. Intraosseous access. Life in the Fast Lane. https://litfl.com/intraosseous-access/. Published April 2019. Accessed June 2019.
- 2.Reid C. iStat Intrasseous. resus.me. http://resus.me/istat-intraosseous/. Published 2014. Accessed June 2019.
- 3.Jousi M, Laukkanen-Nevala P, Nurmi J. Analysing blood from intraosseous access. European Journal of Emergency Medicine. April 2019:77-85. doi:10.1097/mej.0000000000000569
- 4.Tallman C, Darracq M, Young M. Analysis of intraosseous blood samples using an EPOC point of care analyzer during resuscitation. Am J Emerg Med. 2017;35(3):499-501. https://www.ncbi.nlm.nih.gov/pubmed/27998615.
- 5.Jousi M, Saikko S, Nurmi J. Intraosseous blood samples for point-of-care analysis: agreement between intraosseous and arterial analyses. Scand J Trauma Resusc Emerg Med. September 2017. doi:10.1186/s13049-017-0435-4
- 6.Giavarina D. Understanding Bland Altman analysis. Biochem Med. 2015:141-151. doi:10.11613/bm.2015.015
- 7.Reid C. Dogmalysis. Resus:Me. http://resus.me/dogmalysis/. Published 2013. Accessed 2019.